MACHINE LEARNING WITH CALIBRATION TECHNIQUE TO DETECT FRAUDULENT CREDIT CARD TRANSACTIONS¶

Goal:

Predict the probability of an online credit card transaction being fraudulent, based on different properties of the transactions.

Table Of Contents¶

Setup Environment
Data Overview
Optimize Memory
Basic Data Stats
Data Preprocessing for EDA
Exploratory Data Analysis
Statistical Significance Test
Feature Engineering
Dimensionality Reduction
Feature Encoding
Data Preprocessing for Model
Model Building
Evaluation Metrics
Capture Rates and Calibration Curve
LightGBM
Random Forest Classifier
Handling Class Imbalance
Cost Sensitive Learning with Class weights
Model Calibration
Model Tuning
Feature Importance
Partial Dependence and Individual Conditional Expectations
Shap Values

1. Setup Environment ¶

The goal of this section is to:

Import all the packages
Set the options for data visualizations

# Data Manipulation
import numpy as np 
import pandas as pd 

# Data Visualization
import seaborn as sns 
import matplotlib.pyplot as plt
import matplotlib.lines as mlines

# Time
import time
import datetime

# Machine Learning
from   sklearn.preprocessing import LabelEncoder, minmax_scale
from   sklearn.ensemble import RandomForestClassifier
from   sklearn.decomposition import PCA
from   sklearn.model_selection import train_test_split, GridSearchCV
from   sklearn.metrics import confusion_matrix , classification_report, accuracy_score, roc_auc_score, plot_roc_curve, precision_recall_curve, plot_precision_recall_curve
from   sklearn.calibration import calibration_curve
from   sklearn.calibration import CalibratedClassifierCV

from   xgboost import XGBClassifier
from   lightgbm import LGBMClassifier

from   imblearn.over_sampling import RandomOverSampler
from   scipy.stats import chi2_contingency,  f_oneway

import gc
import warnings
from   tqdm import tqdm


# Set Options
pd.set_option('display.max_rows', 800)
pd.set_option('display.max_columns', 500)
%matplotlib inline
warnings.filterwarnings("ignore")

2. Data Overview ¶

Purpose is to:

Load the datasets
Explore the features

The data is broken into two files identity and transaction, which are joined by “TransactionID”.

Note: Not all transactions have corresponding identity information.

Load the transaction and identity datasets using pd.read_csv()

%%time
# Load Data
df_id   = pd.read_csv('Data/train_identity.csv')
df_tran = pd.read_csv('Data/train_transaction.csv')

Wall time: 21.3 s

# Identitiy Data
df_id.sample(6)

Identity Data Description¶

Variables in this table are identity information – network connection information (IP, ISP, Proxy, etc) and digital signature (UA/browser/os/version, etc) associated with transactions. They're collected by Vesta’s fraud protection system and digital security partners. (The field names are masked and pairwise dictionary will not be provided for privacy protection and contract agreement)

Categorical Features:

DeviceType
DeviceInfo
id_12 - id_38

# Transaction Data
df_tran.head()

Transaction Data Description¶

TransactionDT: timedelta from a given reference datetime (not an actual timestamp)
TransactionAMT: transaction payment amount in USD
ProductCD: product code, the product for each transaction
card1 - card6: payment card information, such as card type, card category, issue bank, country, etc.
addr: address
dist: distance
P_ and (R__) emaildomain: purchaser and recipient email domain
C1-C14: counting, such as how many addresses are found to be associated with the payment card, etc. The actual meaning is masked.
D1-D15: timedelta, such as days between previous transaction, etc.
M1-M9: match, such as names on card and address, etc.
Vxxx: Vesta engineered rich features, including ranking, counting, and other entity relations.

3. Optimize Memory Used by Data ¶

There are few features which has data that needs smaller memory size to hold it but the current data type is occupying more memory. Hence reducing memory usage by the data is very much necessary. This section gives a function to modify the data type of each feature.

Memory occupied by the dataframe (in mb)¶

df_id.memory_usage(deep=True).sum() / 1024**2

157.63398933410645

df_tran.memory_usage(deep=True).sum() / 1024**2

2100.701406478882

df_tran.dtypes

TransactionID       int64
isFraud             int64
TransactionDT       int64
TransactionAmt    float64
ProductCD          object
card1               int64
card2             float64
card3             float64
card4              object
card5             float64
card6              object
addr1             float64
addr2             float64
dist1             float64
dist2             float64
P_emaildomain      object
R_emaildomain      object
C1                float64
C2                float64
C3                float64
C4                float64
C5                float64
C6                float64
C7                float64
C8                float64
C9                float64
C10               float64
C11               float64
C12               float64
C13               float64
C14               float64
D1                float64
D2                float64
D3                float64
D4                float64
D5                float64
D6                float64
D7                float64
D8                float64
D9                float64
D10               float64
D11               float64
D12               float64
D13               float64
D14               float64
D15               float64
M1                 object
M2                 object
M3                 object
M4                 object
M5                 object
M6                 object
M7                 object
M8                 object
M9                 object
V1                float64
V2                float64
V3                float64
V4                float64
V5                float64
V6                float64
V7                float64
V8                float64
V9                float64
V10               float64
V11               float64
V12               float64
V13               float64
V14               float64
V15               float64
V16               float64
V17               float64
V18               float64
V19               float64
V20               float64
V21               float64
V22               float64
V23               float64
V24               float64
V25               float64
V26               float64
V27               float64
V28               float64
V29               float64
V30               float64
V31               float64
V32               float64
V33               float64
V34               float64
V35               float64
V36               float64
V37               float64
V38               float64
V39               float64
V40               float64
V41               float64
V42               float64
V43               float64
V44               float64
V45               float64
V46               float64
V47               float64
V48               float64
V49               float64
V50               float64
V51               float64
V52               float64
V53               float64
V54               float64
V55               float64
V56               float64
V57               float64
V58               float64
V59               float64
V60               float64
V61               float64
V62               float64
V63               float64
V64               float64
V65               float64
V66               float64
V67               float64
V68               float64
V69               float64
V70               float64
V71               float64
V72               float64
V73               float64
V74               float64
V75               float64
V76               float64
V77               float64
V78               float64
V79               float64
V80               float64
V81               float64
V82               float64
V83               float64
V84               float64
V85               float64
V86               float64
V87               float64
V88               float64
V89               float64
V90               float64
V91               float64
V92               float64
V93               float64
V94               float64
V95               float64
V96               float64
V97               float64
V98               float64
V99               float64
V100              float64
V101              float64
V102              float64
V103              float64
V104              float64
V105              float64
V106              float64
V107              float64
V108              float64
V109              float64
V110              float64
V111              float64
V112              float64
V113              float64
V114              float64
V115              float64
V116              float64
V117              float64
V118              float64
V119              float64
V120              float64
V121              float64
V122              float64
V123              float64
V124              float64
V125              float64
V126              float64
V127              float64
V128              float64
V129              float64
V130              float64
V131              float64
V132              float64
V133              float64
V134              float64
V135              float64
V136              float64
V137              float64
V138              float64
V139              float64
V140              float64
V141              float64
V142              float64
V143              float64
V144              float64
V145              float64
V146              float64
V147              float64
V148              float64
V149              float64
V150              float64
V151              float64
V152              float64
V153              float64
V154              float64
V155              float64
V156              float64
V157              float64
V158              float64
V159              float64
V160              float64
V161              float64
V162              float64
V163              float64
V164              float64
V165              float64
V166              float64
V167              float64
V168              float64
V169              float64
V170              float64
V171              float64
V172              float64
V173              float64
V174              float64
V175              float64
V176              float64
V177              float64
V178              float64
V179              float64
V180              float64
V181              float64
V182              float64
V183              float64
V184              float64
V185              float64
V186              float64
V187              float64
V188              float64
V189              float64
V190              float64
V191              float64
V192              float64
V193              float64
V194              float64
V195              float64
V196              float64
V197              float64
V198              float64
V199              float64
V200              float64
V201              float64
V202              float64
V203              float64
V204              float64
V205              float64
V206              float64
V207              float64
V208              float64
V209              float64
V210              float64
V211              float64
V212              float64
V213              float64
V214              float64
V215              float64
V216              float64
V217              float64
V218              float64
V219              float64
V220              float64
V221              float64
V222              float64
V223              float64
V224              float64
V225              float64
V226              float64
V227              float64
V228              float64
V229              float64
V230              float64
V231              float64
V232              float64
V233              float64
V234              float64
V235              float64
V236              float64
V237              float64
V238              float64
V239              float64
V240              float64
V241              float64
V242              float64
V243              float64
V244              float64
V245              float64
V246              float64
V247              float64
V248              float64
V249              float64
V250              float64
V251              float64
V252              float64
V253              float64
V254              float64
V255              float64
V256              float64
V257              float64
V258              float64
V259              float64
V260              float64
V261              float64
V262              float64
V263              float64
V264              float64
V265              float64
V266              float64
V267              float64
V268              float64
V269              float64
V270              float64
V271              float64
V272              float64
V273              float64
V274              float64
V275              float64
V276              float64
V277              float64
V278              float64
V279              float64
V280              float64
V281              float64
V282              float64
V283              float64
V284              float64
V285              float64
V286              float64
V287              float64
V288              float64
V289              float64
V290              float64
V291              float64
V292              float64
V293              float64
V294              float64
V295              float64
V296              float64
V297              float64
V298              float64
V299              float64
V300              float64
V301              float64
V302              float64
V303              float64
V304              float64
V305              float64
V306              float64
V307              float64
V308              float64
V309              float64
V310              float64
V311              float64
V312              float64
V313              float64
V314              float64
V315              float64
V316              float64
V317              float64
V318              float64
V319              float64
V320              float64
V321              float64
V322              float64
V323              float64
V324              float64
V325              float64
V326              float64
V327              float64
V328              float64
V329              float64
V330              float64
V331              float64
V332              float64
V333              float64
V334              float64
V335              float64
V336              float64
V337              float64
V338              float64
V339              float64
dtype: object

Certain features occupy more memory than what is needed to store them. Reducing the memory usage by changing data type will speed up the computations.

Let's create a function for that:

int8 / uint8 : consumes 1 byte of memory, range between -128/127 or 0/255
bool : consumes 1 byte, true or false
float16 / int16 / uint16: consumes 2 bytes of memory, range between -32768 and 32767 or 0/65535
float32 / int32 / uint32 : consumes 4 bytes of memory, range between -2147483648 and 2147483647
float64 / int64 / uint64: consumes 8 bytes of memory

print('int64 min: ', np.iinfo(np.int64).min)
print('int64 max: ', np.iinfo(np.int64).max)

int64 min:  -9223372036854775808
int64 max:  9223372036854775807

print('int8 min: ', np.iinfo(np.int8).min)
print('int8 max: ', np.iinfo(np.int8).max)

int8 min:  -128
int8 max:  127

# Reduce memory usage
def reduce_mem_usage(df, verbose=True):
    numerics = ['int16', 'int32', 'int64', 'float16', 'float32', 'float64']
    start_mem = df.memory_usage(deep=True).sum() / 1024**2    
    for col in df.columns:
        col_type = df[col].dtypes
        if col_type in numerics:
            c_min = df[col].min()
            c_max = df[col].max()
            if str(col_type)[:3] == 'int':
                if c_min >= np.iinfo(np.int8).min and c_max <= np.iinfo(np.int8).max:
                    df[col] = df[col].astype(np.int8)
                elif c_min >= np.iinfo(np.int16).min and c_max <= np.iinfo(np.int16).max:
                    df[col] = df[col].astype(np.int16)
                elif c_min >= np.iinfo(np.int32).min and c_max <= np.iinfo(np.int32).max:
                    df[col] = df[col].astype(np.int32)
                elif c_min >= np.iinfo(np.int64).min and c_max <= np.iinfo(np.int64).max:
                    df[col] = df[col].astype(np.int64)  
            else:
                if c_min >= np.finfo(np.float16).min and c_max <= np.finfo(np.float16).max:
                    df[col] = df[col].astype(np.float16)
                elif c_min >= np.finfo(np.float32).min and c_max <= np.finfo(np.float32).max:
                    df[col] = df[col].astype(np.float32)
                else:
                    df[col] = df[col].astype(np.float64)    
    end_mem = df.memory_usage(deep=True).sum() / 1024**2
    if verbose: print('Mem. usage decreased to {:5.2f} Mb ({:.1f}% reduction)'.format(end_mem, 100 * (start_mem - end_mem) / start_mem))
    return df

Use the defined function to reduce the memory usage

# Reduce the memory size of the dataframe
df_id   = reduce_mem_usage(df_id)
df_tran = reduce_mem_usage(df_tran)

Mem. usage decreased to 138.38 Mb (12.2% reduction)
Mem. usage decreased to 867.89 Mb (58.7% reduction)

4. Basic Data Stats ¶

Before attempting to solve the problem, it's very important to have a good understanding of data.

The goal of this section is to:

Get the dimensions of data
Get the summary of data
Get various statistics of data

Shape of dataframe¶

# Dimensions of identity dataset
print(df_id.shape)

(144233, 41)

The dataset has 144233 rows and 41 columns

# Dimensions of transaction dataset
print(df_tran.shape)

(590540, 394)

The dataset has 590540 rows and 394 columns

Check how many transactions has ID info

# How many had ID info?
df_tran.TransactionID.isin(df_id.TransactionID).sum()

144233

Summary of dataframe¶

df_id.head()

from pandas_summary import DataFrameSummary
df_id_summary = DataFrameSummary(df_id)
df_id_summary.summary()

By looking at the summary of datasets, it's clear there is a lot of missing values in the dataset.

Let's get missing value stats and various other stats of columns in dataframe.

Stats on Transaction Dataset¶

from pandas_summary import DataFrameSummary
df_tran_summary = DataFrameSummary(df_tran)
df_tran_summary.summary()

Check class imbalance

df_tran.loc[:, 'isFraud'].value_counts()

0    569877
1     20663
Name: isFraud, dtype: int64

Inferences:

Lot of interesting things can be observed here:

Rows in identity dataset are less than transaction dataset, that means only a subset of transactions in transactions dataset has identity data
Both datasets have the common and unique key as TransactionID, both can be joined at this unique key
id_24, id_25, dist2, D7 and many more columns have 90%+ missing values, which means that these columns are probably useless so need to drop it for now
Columns from V1 to V339 in transaction dataset are numeric whereas columns from id_01 to id_39 are of mixed datatype
TransactionDT column is a timedelta from a given reference datetime (not an actual timestamp). But reference datetime is not known, so need to assume it and convert it to date format
Target class is imbalanced. So no need to drop the columns where one category contains the majority of rows

5. Data Preprocessing for EDA ¶

The goal of this section is to:

Merge two datasets
Drop the columns based on the inferences from previous section
Create date features from transaction date

Let's start with the first task to merge datasets to form one.

Merge the datasets¶

# Merge transaction dataset and identity dataset 
df = df_tran.merge(df_id, how='left', left_index=True, right_index=True, on='TransactionID')

del df_tran, df_id

gc.collect()

0

Get dimensions of training dataset

# Dimentions of data
df.shape

(590540, 434)

Since left join was performed on transaction dataset, number of rows are same as transaction dataset.

Add missing flag¶

# Add flag column for missing values
for col in df.columns:
    df[col+"_missing_flag"] = df[col].isnull()
    
df.head()

Clean Data¶

Let's drop the columns which may not be useful for our analysis

Create a missing value flag column for the columns we are dropping which have more than 90% missing values, there might be some specific pattern associated with missing values and transaction being fraud

# Drop the columns where one category contains more than 90% values
drop_cols = []

for col in df.columns:
    missing_share = df[col].isnull().sum()/df.shape[0]
    if missing_share > 0.9:
        drop_cols.append(col)
        print(col)
        # df[col + "_missing_flag"] = df[col].isnull()
    
good_cols = [col for col in df.columns if col not in drop_cols]

dist2
D7
id_07
id_08
id_18
id_21
id_22
id_23
id_24
id_25
id_26
id_27

Remove the columns which doesn't having any variance

# Drop the columns which have only one unique value
drop_cols = []
for col in good_cols:
    unique_value = df[col].nunique()
    if unique_value == 1:
        drop_cols.append(col)
        print(col)
good_cols = [col for col in good_cols if col not in drop_cols]

TransactionID_missing_flag
isFraud_missing_flag
TransactionDT_missing_flag
TransactionAmt_missing_flag
ProductCD_missing_flag
card1_missing_flag
C1_missing_flag
C2_missing_flag
C3_missing_flag
C4_missing_flag
C5_missing_flag
C6_missing_flag
C7_missing_flag
C8_missing_flag
C9_missing_flag
C10_missing_flag
C11_missing_flag
C12_missing_flag
C13_missing_flag
C14_missing_flag

Filter the dataset with only good columns

# Filter the data for relevant columns only
df = df[good_cols]

Get dimentions of training dataset

# Dimentions of data
df.shape

(590540, 836)

Create date features¶

Let's create date features from TransactionDT features

# Date features
START_DATE         = '2017-12-01'
startdate          = datetime.datetime.strptime(START_DATE, "%Y-%m-%d")
df["Date"]         = df['TransactionDT'].apply(lambda x: (startdate + datetime.timedelta(seconds=x)))

df['_Weekdays']    = df['Date'].dt.dayofweek
df['_Hours']       = df['Date'].dt.hour
df['_Days']        = df['Date'].dt.day

df = reduce_mem_usage(df)

Mem. usage decreased to 1449.38 Mb (0.8% reduction)

6. Exploratory Data Analysis ¶

Exploratory data analysis is an approach to analyze or investigate data sets to find out patterns and see if any of the variables can be useful to explain / predict the y variables.

Visual methods are often used to summarise the data. Primarily EDA is for seeing what the data can tell us beyond the formal modeling or hypothesis testing tasks.

The goal of this section is to:

Check if the target variable is balanced or is there a need to balance the target variable
Analyze the transaction amount
Get insights or relationships from the data which would be useful from business perspective.

Check distribution of target variable¶

# Get count of target class
df['isFraud'].value_counts()

0    569877
1     20663
Name: isFraud, dtype: int64

Let's check the distribution of target class using a bar plot and check the proportion of transactions amounts being fraud

# Draw a countplot to check the distribution of target variable
df['TransactionAmt'] = df['TransactionAmt'].astype(float)
total = len(df)
total_amt = df.groupby(['isFraud'])['TransactionAmt'].sum().sum()
plt.figure(figsize=(16,6))

plt.subplot(121)
g = sns.countplot(x='isFraud', data=df )
g.set_title("Fraud Transactions Distribution \n 0: No Fraud | 1: Fraud", fontsize=18)
g.set_xlabel("Is fraud?", fontsize=18)
g.set_ylabel('Count', fontsize=18)
for p in g.patches:
    height = p.get_height()
    g.text(p.get_x()+p.get_width()/2.,
            height + 3,
            '{:1.2f}%'.format(height/total*100),
            ha="center", fontsize=15) 

perc_amt = (df.groupby(['isFraud'])['TransactionAmt'].sum())
perc_amt = perc_amt.reset_index()

plt.subplot(122)
g1 = sns.barplot(x='isFraud', y='TransactionAmt',  dodge=True, data=perc_amt)
g1.set_title("% Total Amount in Transaction Amt \n 0: No Fraud | 1: Fraud", fontsize=18)
g1.set_xlabel("Is fraud?", fontsize=18)
g1.set_ylabel('Total Transaction Amount Scalar', fontsize=18)
for p in g1.patches:
    height = p.get_height()
    g1.text(p.get_x()+p.get_width()/2.,
            height + 3,
            '{:1.2f}%'.format(height/total_amt * 100),
            ha="center", fontsize=15) 
    
plt.show()

# Average transaction amount by Y
df.groupby('isFraud')['TransactionAmt'].mean()

isFraud
0    134.511857
1    149.244353
Name: TransactionAmt, dtype: float64

Inferences:

The target variable is imbalanced. 3.5% transactions are Fraud
Around same % of transaction amounts are fraud

Let's explore the Transaction amount further

Check distribution of Transaction Amount¶

# Distribution plot of Transaction Amount
plt.figure(figsize=(16,12))

sns.distplot(df['TransactionAmt'])
plt.title("Transaction Amount Distribution",fontsize=18)
plt.ylabel("Probability")

Text(0, 0.5, 'Probability')

There are certain transactions which are of very high amount, let's remove those to check the distribution

# Distribution plot of Transaction Amount less than 1000
plt.figure(figsize=(16,12))

plt.suptitle('Transaction Values Distribution', fontsize=22)
sns.distplot(df[df['TransactionAmt'] <= 1000]['TransactionAmt'])
plt.title("Transaction Amount Distribuition <= 1000", fontsize=18)
plt.xlabel("Transaction Amount", fontsize=15)
plt.ylabel("Probability", fontsize=15)

plt.show()

Most transactions lie in < $200 range

Transaction amount is right skewed.

Let's look at the log of transaction amount

# Distribution plot of Transaction Amount less than 1000
plt.figure(figsize=(16,12))

plt.suptitle('Transaction Values Distribution', fontsize=22)
sns.distplot(np.log(df['TransactionAmt']))
plt.title("Transaction Amount (Log) Distribuition", fontsize=18)
plt.xlabel("Transaction Amount", fontsize=15)
plt.ylabel("Probability", fontsize=15)

plt.show()

Inferences:

Transaction Amount is right skewed.
Log of transaction amount is almost normally distributed, so use log of transaction amount while building the model

Product Features¶

Distribution of ProductCD
Distribution of Frauds by Product

def plot_cat_feat_dist(df, col):
    tmp = pd.crosstab(df[col], df['isFraud'], normalize='index') * 100
    tmp = tmp.reset_index()
    tmp.rename(columns={0:'NoFraud', 1:'Fraud'}, inplace=True)

    plt.figure(figsize=(16,12))
    plt.suptitle(f'{col} Distributions', fontsize=22)

    plt.subplot(221)
    g = sns.countplot(x=col, data=df, order=tmp[col].values)

    g.set_title(f"{col} Distribution", fontsize=16)
    g.set_xlabel(f"{col} Name", fontsize=17)
    g.set_ylabel("Count", fontsize=17)
    for p in g.patches:
        height = p.get_height()
        g.text(p.get_x()+p.get_width()/2.,
                height + 3,
                '{:1.2f}%'.format(height/total*100),
                ha="center", fontsize=14) 

    plt.subplot(222)
    g1 = sns.countplot(x=col, hue='isFraud', data=df, order=tmp[col].values)
    plt.legend(title='Fraud', loc='best', labels=['No', 'Yes'])
    gt = g1.twinx()
    gt = sns.pointplot(x=col, y='Fraud', data=tmp, color='black', order=tmp[col].values, legend=False)
    gt.set_ylabel("% of Fraud Transactions", fontsize=16)

    g1.set_title(f"{col} Distribution by Target Variable (isFraud) ", fontsize=16)
    g1.set_xlabel(f"{col} Name", fontsize=17)
    g1.set_ylabel("Count", fontsize=17)


    plt.subplots_adjust(hspace = 0.4, top = 0.85)

    plt.show()

plot_cat_feat_dist(df, "ProductCD")

# Average fraud per transaction by ProductCD
df.groupby('ProductCD')['isFraud'].mean()

ProductCD
C    0.116873
H    0.047662
R    0.037826
S    0.058996
W    0.020399
Name: isFraud, dtype: float64

Inferences:

75% of the transactions are for Product Catergory W
11.6% of the transactions are for Product Category C
Fraud Transaction rate is maximum for Product Category C and minimum for Product Category W

Card Features¶

# Card 4
plot_cat_feat_dist(df, "card4")

# Average fraud per transaction by Card4
df.groupby('card4')['isFraud'].mean()

card4
american express    0.028698
discover            0.077282
mastercard          0.034331
visa                0.034756
Name: isFraud, dtype: float64

Inferences:

97% of transactions are from Mastercard(32%) and Visa(65%
Fraud transaction rate is highest for discover cards(~8%) against ~3.5% of Mastercard and Visa and 2.87% in American Express

# Card 6
plot_cat_feat_dist(df, "card6")

# Average fraud per transaction by Card6
df.groupby('card6')['isFraud'].mean()

card6
charge card        0.000000
credit             0.066785
debit              0.024263
debit or credit    0.000000
Name: isFraud, dtype: float64

Inferences:

Almost all the transactions are from Credit and Debit cards.
Debit card transactions are almost 3 times as compared to credit card transactions.
Fraud transaction rate is high for Credit cards as compared to Debit cards.

P emaildomain¶

It has multiple domains, let's group them by the respective enterprises
Set all values with less than 500 entries as "Others"

df.loc[df['P_emaildomain'].isin(['gmail.com', 'gmail']),'P_emaildomain'] = 'Google'

df.loc[df['P_emaildomain'].isin(['yahoo.com', 'yahoo.com.mx',  'yahoo.co.uk',
                                         'yahoo.co.jp', 'yahoo.de', 'yahoo.fr',
                                         'yahoo.es']), 'P_emaildomain'] = 'Yahoo Mail'
df.loc[df['P_emaildomain'].isin(['hotmail.com','outlook.com','msn.com', 'live.com.mx', 
                                         'hotmail.es','hotmail.co.uk', 'hotmail.de',
                                         'outlook.es', 'live.com', 'live.fr',
                                         'hotmail.fr']), 'P_emaildomain'] = 'Microsoft'
df.loc[df.P_emaildomain.isin(df.P_emaildomain\
                                         .value_counts()[df.P_emaildomain.value_counts() <= 500 ]\
                                         .index), 'P_emaildomain'] = "Others"
df.P_emaildomain.fillna("NoInf", inplace=True)

def plot_cat_with_amt(df, col, lim=2000):
    tmp = pd.crosstab(df[col], df['isFraud'], normalize='index') * 100
    tmp = tmp.reset_index()
    tmp.rename(columns={0:'NoFraud', 1:'Fraud'}, inplace=True)
    
    plt.figure(figsize=(16,14))    
    plt.suptitle(f'{col} Distributions ', fontsize=24)
    
    plt.subplot(211)
    g = sns.countplot( x=col,  data=df, order=list(tmp[col].values))
    gt = g.twinx()
    gt = sns.pointplot(x=col, y='Fraud', data=tmp, order=list(tmp[col].values),
                       color='black', legend=False, )
    gt.set_ylim(0,tmp['Fraud'].max()*1.1)
    gt.set_ylabel("%Fraud Transactions", fontsize=16)
    g.set_title(f"Share of {col} categories and % of Fraud Transactions", fontsize=18)
    g.set_xlabel(f"{col} Category Names", fontsize=16)
    g.set_ylabel("Count", fontsize=17)
    g.set_xticklabels(g.get_xticklabels(),rotation=45)
    sizes = []
    for p in g.patches:
        height = p.get_height()
        sizes.append(height)
        g.text(p.get_x()+p.get_width()/2.,
                height + 3,
                '{:1.2f}%'.format(height/total*100),
                ha="center",fontsize=12) 
        
    g.set_ylim(0,max(sizes)*1.15)
    
    #########################################################################
    perc_amt = (df.groupby(['isFraud',col])['TransactionAmt'].sum() \
                / df.groupby([col])['TransactionAmt'].sum() * 100).unstack('isFraud')
    perc_amt = perc_amt.reset_index()
    perc_amt.rename(columns={0:'NoFraud', 1:'Fraud'}, inplace=True)
    amt = df.groupby([col])['TransactionAmt'].sum().reset_index()
    perc_amt = perc_amt.fillna(0)
    plt.subplot(212)
    g1 = sns.barplot(x=col, y='TransactionAmt', 
                       data=amt, 
                       order=list(tmp[col].values))
    g1t = g1.twinx()
    g1t = sns.pointplot(x=col, y='Fraud', data=perc_amt, 
                        order=list(tmp[col].values),
                       color='black', legend=False, )
    g1t.set_ylim(0,perc_amt['Fraud'].max()*1.1)
    g1t.set_ylabel("%Fraud Total Amount", fontsize=16)
    g.set_xticklabels(g.get_xticklabels(),rotation=45)
    g1.set_title(f"Transactions amount by {col} categories and % of Fraud Transactions (Amounts)", fontsize=18)
    g1.set_xlabel(f"{col} Category Names", fontsize=16)
    g1.set_ylabel("Transaction Total Amount(U$)", fontsize=16)
    g1.set_xticklabels(g.get_xticklabels(),rotation=45)    
    
    for p in g1.patches:
        height = p.get_height()
        g1.text(p.get_x()+p.get_width()/2.,
                height + 3,
                '{:1.2f}%'.format(height/total_amt*100),
                ha="center",fontsize=12) 
        
    plt.subplots_adjust(hspace=.4, top = 0.9)
    plt.show()

plot_cat_with_amt(df, 'P_emaildomain')

# Average fraud per transaction by Card6
df.groupby('P_emaildomain')['isFraud'].mean()

P_emaildomain
aim.com             0.126984
anonymous.com       0.023217
aol.com             0.021811
att.net             0.007439
bellsouth.net       0.027763
cableone.net        0.018868
centurylink.net     0.000000
cfl.rr.com          0.000000
charter.net         0.030637
comcast.net         0.031187
cox.net             0.020818
earthlink.net       0.021401
embarqmail.com      0.034615
frontier.com        0.028571
frontiernet.net     0.025641
gmail               0.022177
gmail.com           0.043542
gmx.de              0.000000
hotmail.co.uk       0.000000
hotmail.com         0.052950
hotmail.de          0.000000
hotmail.es          0.065574
hotmail.fr          0.000000
icloud.com          0.031434
juno.com            0.018634
live.com            0.027622
live.com.mx         0.054740
live.fr             0.000000
mac.com             0.032110
mail.com            0.189624
me.com              0.017740
msn.com             0.021994
netzero.com         0.000000
netzero.net         0.005102
optonline.net       0.016815
outlook.com         0.094584
outlook.es          0.130137
prodigy.net.mx      0.004831
protonmail.com      0.407895
ptd.net             0.000000
q.com               0.000000
roadrunner.com      0.009836
rocketmail.com      0.003012
sbcglobal.net       0.004040
sc.rr.com           0.006098
servicios-ta.com    0.000000
suddenlink.net      0.022857
twc.com             0.000000
verizon.net         0.008133
web.de              0.000000
windstream.net      0.000000
yahoo.co.jp         0.000000
yahoo.co.uk         0.000000
yahoo.com           0.022757
yahoo.com.mx        0.010369
yahoo.de            0.000000
yahoo.es            0.014925
yahoo.fr            0.034965
ymail.com           0.020868
Name: isFraud, dtype: float64

Inferences:

Majority of transactions are with P_emaildomain as Google, Microsoft and Yahoo Mail
There isn't any information about P_emaildomain of around 16% transactions in terms of count and 14.11% in terms of amount
Fraud transaction rate for Microsoft is high as compared to Google and Yahoo mail
Fraud transaction rate (amount) for Google is high as comapred to Microsoft and Yahoo mail

R-Email Domain¶

It has multiple domains, let's group them by the respective enterprises
Set all values with less than 500 entries as "Others"

df.loc[df['R_emaildomain'].isin(['gmail.com', 'gmail']),'R_emaildomain'] = 'Google'

df.loc[df['R_emaildomain'].isin(['yahoo.com', 'yahoo.com.mx',  'yahoo.co.uk',
                                             'yahoo.co.jp', 'yahoo.de', 'yahoo.fr',
                                             'yahoo.es']), 'R_emaildomain'] = 'Yahoo Mail'
df.loc[df['R_emaildomain'].isin(['hotmail.com','outlook.com','msn.com', 'live.com.mx', 
                                             'hotmail.es','hotmail.co.uk', 'hotmail.de',
                                             'outlook.es', 'live.com', 'live.fr',
                                             'hotmail.fr']), 'R_emaildomain'] = 'Microsoft'
df.loc[df.R_emaildomain.isin(df.R_emaildomain\
                                         .value_counts()[df.R_emaildomain.value_counts() <= 300 ]\
                                         .index), 'R_emaildomain'] = "Others"
df.R_emaildomain.fillna("NoInf", inplace=True)

plot_cat_with_amt(df, 'R_emaildomain')

# Average fraud per transaction by Card6
df.groupby('R_emaildomain')['isFraud'].mean()

R_emaildomain
aim.com             0.027778
anonymous.com       0.029130
aol.com             0.034855
att.net             0.000000
bellsouth.net       0.004739
cableone.net        0.000000
centurylink.net     0.000000
cfl.rr.com          0.000000
charter.net         0.039370
comcast.net         0.011589
cox.net             0.023965
earthlink.net       0.075949
embarqmail.com      0.000000
frontier.com        0.000000
frontiernet.net     0.000000
gmail               0.000000
gmail.com           0.119184
gmx.de              0.000000
hotmail.co.uk       0.000000
hotmail.com         0.077793
hotmail.de          0.000000
hotmail.es          0.068493
hotmail.fr          0.000000
icloud.com          0.128755
juno.com            0.000000
live.com            0.049869
live.com.mx         0.058355
live.fr             0.000000
mac.com             0.009174
mail.com            0.377049
me.com              0.019784
msn.com             0.001174
netzero.com         0.000000
netzero.net         0.222222
optonline.net       0.010695
outlook.com         0.165138
outlook.es          0.131640
prodigy.net.mx      0.004831
protonmail.com      0.951220
ptd.net             0.000000
q.com               0.000000
roadrunner.com      0.000000
rocketmail.com      0.043478
sbcglobal.net       0.001812
sc.rr.com           0.000000
scranton.edu        0.000000
servicios-ta.com    0.000000
suddenlink.net      0.040000
twc.com             0.000000
verizon.net         0.000000
web.de              0.000000
windstream.net      0.000000
yahoo.co.jp         0.000000
yahoo.co.uk         0.000000
yahoo.com           0.051512
yahoo.com.mx        0.010610
yahoo.de            0.000000
yahoo.es            0.035088
yahoo.fr            0.036496
ymail.com           0.038647
Name: isFraud, dtype: float64

Inferences:

There isn't any information about R_emaildomain for Majority of transactions (76.75% count , 85.62% amount)
Fraud transaction rate for Google is high as compared to Yahoo, anaonymous.com and Microsoft

Days of the Month¶

Reference date is not known, it has been assumed. So can't say concretely if the day number is correct

plot_cat_with_amt(df, '_Days')

Inferences:

The perc of fraud transactions is highest towards the beginning and the end of the month. Might be accelerated at the time of receiving pay-checks.
Incidentally, fraud transaction rate is high on the days when number of transactions are less
Day 29,30 and 31 are having less transactions, looks like people are cautious with spending in those times.

Days of the week¶

Reference date is not known, it has been assumed. So can't say concretely if the day number is correct

plot_cat_with_amt(df, '_Weekdays')

Inferences:

Surprisingly fraud transaction rate is high on the days when number of transactions and transaction amounts are less. Day 0 and 6
Day 0 and 6 have less transactions, these might be weekend days

Hour of the Day¶

plot_cat_with_amt(df, '_Hours')

Inferences:

Transactions start decreasing mid night but the fraud rate starts increasing
Transactions from 3 AM to 12 PM needs to monitored very closely

Device Type¶

plot_cat_with_amt(df, "DeviceType")

Inferences:

Device type is known for only 24% of the transactions
Due to lack of data points, we can't infer from this analysis

Columns from identity data¶

for col in ['id_12', 'id_15', 'id_16', 'id_28', 'id_29']:
    df[col] = df[col].fillna('NaN')
    plot_cat_with_amt(df, col)

df.loc[df['id_30'].str.contains('Windows', na=False), 'id_30'] = 'Windows'
df.loc[df['id_30'].str.contains('iOS', na=False), 'id_30'] = 'iOS'
df.loc[df['id_30'].str.contains('Mac OS', na=False), 'id_30'] = 'Mac'
df.loc[df['id_30'].str.contains('Android', na=False), 'id_30'] = 'Android'
df['id_30'].fillna("NAN", inplace=True)

plot_cat_with_amt(df, "id_30")

df.loc[df['id_31'].str.contains('chrome', na=False), 'id_31'] = 'Chrome'
df.loc[df['id_31'].str.contains('firefox', na=False), 'id_31'] = 'Firefox'
df.loc[df['id_31'].str.contains('safari', na=False), 'id_31'] = 'Safari'
df.loc[df['id_31'].str.contains('edge', na=False), 'id_31'] = 'Edge'
df.loc[df['id_31'].str.contains('ie', na=False), 'id_31'] = 'IE'
df.loc[df['id_31'].str.contains('samsung', na=False), 'id_31'] = 'Samsung'
df.loc[df['id_31'].str.contains('opera', na=False), 'id_31'] = 'Opera'
df['id_31'].fillna("NAN", inplace=True)
df.loc[df.id_31.isin(df.id_31.value_counts()[df.id_31.value_counts() < 200].index), 'id_31'] = "Others"
plot_cat_with_amt(df, "id_31")

Get column names¶

cat_columns = df.select_dtypes(include=['object']).columns
len(cat_columns)

29

binary_columns = [col for col in df.columns if df[col].nunique() == 2]
len(binary_columns)

435

num_columns = [col for col in df.columns if (col not in cat_columns) & (col not in binary_columns)]
len(num_columns)

389

cat_columns = cat_columns.to_list() + binary_columns

7. Statistical Significance test ¶

Chi square test for categorical columns¶

from   scipy.stats import chi2_contingency

# significance value
alpha = 0.05

significant_categorical_variables = []

for col in cat_columns:  
    # Create a crosstab table
    temp = pd.crosstab(df[col],df['isFraud'].astype('category'))
    
    # Get chi-square value , p-value, degrees of freedom, expected frequencies using the function chi2_contingency
    stat, p, dof, expected = chi2_contingency(temp)
    
    # Determine whether to reject or keep your null hypothesis
    print(col.ljust(40), ',  chisquared=%.5f,   p-value=%.5f' % (stat, p))
    if p <= alpha:
        significant_categorical_variables.append(col)
    else:
        ""

ProductCD                                ,  chisquared=16742.17153,   p-value=0.00000
card4                                    ,  chisquared=364.87414,   p-value=0.00000
card6                                    ,  chisquared=5957.03229,   p-value=0.00000
P_emaildomain                            ,  chisquared=3497.81283,   p-value=0.00000
R_emaildomain                            ,  chisquared=17297.50859,   p-value=0.00000
M1                                       ,  chisquared=0.00003,   p-value=0.99581
M2                                       ,  chisquared=438.61321,   p-value=0.00000
M3                                       ,  chisquared=477.66057,   p-value=0.00000
M4                                       ,  chisquared=6450.44798,   p-value=0.00000
M5                                       ,  chisquared=242.42169,   p-value=0.00000
M6                                       ,  chisquared=227.96414,   p-value=0.00000
M7                                       ,  chisquared=11.25610,   p-value=0.00079
M8                                       ,  chisquared=88.53022,   p-value=0.00000
M9                                       ,  chisquared=250.37250,   p-value=0.00000
id_12                                    ,  chisquared=429.84996,   p-value=0.00000
id_15                                    ,  chisquared=421.13420,   p-value=0.00000
id_16                                    ,  chisquared=366.45700,   p-value=0.00000
id_28                                    ,  chisquared=420.29657,   p-value=0.00000
id_29                                    ,  chisquared=420.22938,   p-value=0.00000
id_30                                    ,  chisquared=176.98386,   p-value=0.00000
id_31                                    ,  chisquared=424.58675,   p-value=0.00000
id_33                                    ,  chisquared=212.82695,   p-value=0.98357
id_34                                    ,  chisquared=11.47415,   p-value=0.00942
id_35                                    ,  chisquared=2.26392,   p-value=0.13242
id_36                                    ,  chisquared=0.02303,   p-value=0.87939
id_37                                    ,  chisquared=1.24550,   p-value=0.26441
id_38                                    ,  chisquared=2.35329,   p-value=0.12502
DeviceType                               ,  chisquared=0.39659,   p-value=0.52885
DeviceInfo                               ,  chisquared=1476.08487,   p-value=1.00000
isFraud                                  ,  chisquared=590510.38453,   p-value=0.00000
M1                                       ,  chisquared=0.00003,   p-value=0.99581
M2                                       ,  chisquared=438.61321,   p-value=0.00000
M3                                       ,  chisquared=477.66057,   p-value=0.00000
M5                                       ,  chisquared=242.42169,   p-value=0.00000
M6                                       ,  chisquared=227.96414,   p-value=0.00000
M7                                       ,  chisquared=11.25610,   p-value=0.00079
M8                                       ,  chisquared=88.53022,   p-value=0.00000
M9                                       ,  chisquared=250.37250,   p-value=0.00000
V1                                       ,  chisquared=0.08480,   p-value=0.77090
V14                                      ,  chisquared=1.85823,   p-value=0.17283
V41                                      ,  chisquared=6.45761,   p-value=0.01105
V65                                      ,  chisquared=2.95009,   p-value=0.08587
V88                                      ,  chisquared=0.06115,   p-value=0.80468
V107                                     ,  chisquared=3.20035,   p-value=0.07362
V305                                     ,  chisquared=0.95990,   p-value=0.32721
id_35                                    ,  chisquared=2.26392,   p-value=0.13242
id_36                                    ,  chisquared=0.02303,   p-value=0.87939
id_37                                    ,  chisquared=1.24550,   p-value=0.26441
id_38                                    ,  chisquared=2.35329,   p-value=0.12502
DeviceType                               ,  chisquared=0.39659,   p-value=0.52885
card2_missing_flag                       ,  chisquared=40.68296,   p-value=0.00000
card3_missing_flag                       ,  chisquared=4.41810,   p-value=0.03556
card4_missing_flag                       ,  chisquared=3.52353,   p-value=0.06050
card5_missing_flag                       ,  chisquared=25.61819,   p-value=0.00000
card6_missing_flag                       ,  chisquared=4.52321,   p-value=0.03344
addr1_missing_flag                       ,  chisquared=15016.72347,   p-value=0.00000
addr2_missing_flag                       ,  chisquared=15016.72347,   p-value=0.00000
dist1_missing_flag                       ,  chisquared=2672.80512,   p-value=0.00000
dist2_missing_flag                       ,  chisquared=4898.54088,   p-value=0.00000
P_emaildomain_missing_flag               ,  chisquared=98.80683,   p-value=0.00000
R_emaildomain_missing_flag               ,  chisquared=11593.85862,   p-value=0.00000
D1_missing_flag                          ,  chisquared=0.02818,   p-value=0.86669
D2_missing_flag                          ,  chisquared=1770.65814,   p-value=0.00000
D3_missing_flag                          ,  chisquared=691.46860,   p-value=0.00000
D4_missing_flag                          ,  chisquared=8.39702,   p-value=0.00376
D5_missing_flag                          ,  chisquared=225.02626,   p-value=0.00000
D6_missing_flag                          ,  chisquared=12282.70285,   p-value=0.00000
D7_missing_flag                          ,  chisquared=15972.26519,   p-value=0.00000
D8_missing_flag                          ,  chisquared=12263.96026,   p-value=0.00000
D9_missing_flag                          ,  chisquared=12263.96026,   p-value=0.00000
D10_missing_flag                         ,  chisquared=671.51048,   p-value=0.00000
D11_missing_flag                         ,  chisquared=4605.06821,   p-value=0.00000
D12_missing_flag                         ,  chisquared=14617.29683,   p-value=0.00000
D13_missing_flag                         ,  chisquared=11641.55968,   p-value=0.00000
D14_missing_flag                         ,  chisquared=13502.69648,   p-value=0.00000
D15_missing_flag                         ,  chisquared=524.34621,   p-value=0.00000
M1_missing_flag                          ,  chisquared=4720.57858,   p-value=0.00000
M2_missing_flag                          ,  chisquared=4720.57858,   p-value=0.00000
M3_missing_flag                          ,  chisquared=4720.57858,   p-value=0.00000
M4_missing_flag                          ,  chisquared=4291.55730,   p-value=0.00000
M5_missing_flag                          ,  chisquared=143.24698,   p-value=0.00000
M6_missing_flag                          ,  chisquared=8958.36720,   p-value=0.00000
M7_missing_flag                          ,  chisquared=2876.27247,   p-value=0.00000
M8_missing_flag                          ,  chisquared=2876.92899,   p-value=0.00000
M9_missing_flag                          ,  chisquared=2876.92899,   p-value=0.00000
V1_missing_flag                          ,  chisquared=4605.06821,   p-value=0.00000
V2_missing_flag                          ,  chisquared=4605.06821,   p-value=0.00000
V3_missing_flag                          ,  chisquared=4605.06821,   p-value=0.00000
V4_missing_flag                          ,  chisquared=4605.06821,   p-value=0.00000
V5_missing_flag                          ,  chisquared=4605.06821,   p-value=0.00000
V6_missing_flag                          ,  chisquared=4605.06821,   p-value=0.00000
V7_missing_flag                          ,  chisquared=4605.06821,   p-value=0.00000
V8_missing_flag                          ,  chisquared=4605.06821,   p-value=0.00000
V9_missing_flag                          ,  chisquared=4605.06821,   p-value=0.00000
V10_missing_flag                         ,  chisquared=4605.06821,   p-value=0.00000
V11_missing_flag                         ,  chisquared=4605.06821,   p-value=0.00000
V12_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V13_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V14_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V15_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V16_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V17_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V18_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V19_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V20_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V21_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V22_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V23_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V24_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V25_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V26_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V27_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V28_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V29_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V30_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V31_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V32_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V33_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V34_missing_flag                         ,  chisquared=670.26786,   p-value=0.00000
V35_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V36_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V37_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V38_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V39_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V40_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V41_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V42_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V43_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V44_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V45_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V46_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V47_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V48_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V49_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V50_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V51_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V52_missing_flag                         ,  chisquared=8.24695,   p-value=0.00408
V53_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V54_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V55_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V56_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V57_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V58_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V59_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V60_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V61_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V62_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V63_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V64_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V65_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V66_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V67_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V68_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V69_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V70_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V71_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V72_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V73_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V74_missing_flag                         ,  chisquared=1448.91324,   p-value=0.00000
V75_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V76_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V77_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V78_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V79_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V80_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V81_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V82_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V83_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V84_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V85_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V86_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V87_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V88_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V89_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V90_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V91_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V92_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V93_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V94_missing_flag                         ,  chisquared=522.48477,   p-value=0.00000
V95_missing_flag                         ,  chisquared=2.86829,   p-value=0.09034
V96_missing_flag                         ,  chisquared=2.86829,   p-value=0.09034
V97_missing_flag                         ,  chisquared=2.86829,   p-value=0.09034
V98_missing_flag                         ,  chisquared=2.86829,   p-value=0.09034
V99_missing_flag                         ,  chisquared=2.86829,   p-value=0.09034
V100_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V101_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V102_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V103_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V104_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V105_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V106_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V107_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V108_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V109_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V110_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V111_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V112_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V113_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V114_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V115_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V116_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V117_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V118_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V119_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V120_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V121_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V122_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V123_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V124_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V125_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V126_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V127_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V128_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V129_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V130_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V131_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V132_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V133_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V134_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V135_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V136_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V137_missing_flag                        ,  chisquared=2.86829,   p-value=0.09034
V138_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V139_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V140_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V141_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V142_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V143_missing_flag                        ,  chisquared=257.28425,   p-value=0.00000
V144_missing_flag                        ,  chisquared=257.28425,   p-value=0.00000
V145_missing_flag                        ,  chisquared=257.28425,   p-value=0.00000
V146_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V147_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V148_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V149_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V150_missing_flag                        ,  chisquared=257.28425,   p-value=0.00000
V151_missing_flag                        ,  chisquared=257.28425,   p-value=0.00000
V152_missing_flag                        ,  chisquared=257.28425,   p-value=0.00000
V153_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V154_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V155_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V156_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V157_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V158_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V159_missing_flag                        ,  chisquared=257.28425,   p-value=0.00000
V160_missing_flag                        ,  chisquared=257.28425,   p-value=0.00000
V161_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V162_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V163_missing_flag                        ,  chisquared=256.78110,   p-value=0.00000
V164_missing_flag                        ,  chisquared=257.28425,   p-value=0.00000
V165_missing_flag                        ,  chisquared=257.28425,   p-value=0.00000
V166_missing_flag                        ,  chisquared=257.28425,   p-value=0.00000
V167_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V168_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V169_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V170_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V171_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V172_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V173_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V174_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V175_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V176_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V177_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V178_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V179_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V180_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V181_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V182_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V183_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V184_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V185_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V186_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V187_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V188_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V189_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V190_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V191_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V192_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V193_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V194_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V195_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V196_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V197_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V198_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V199_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V200_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V201_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V202_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V203_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V204_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V205_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V206_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V207_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V208_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V209_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V210_missing_flag                        ,  chisquared=10589.78839,   p-value=0.00000
V211_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V212_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V213_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V214_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V215_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V216_missing_flag                        ,  chisquared=10515.97912,   p-value=0.00000
V217_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V218_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V219_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V220_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V221_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V222_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V223_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V224_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V225_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V226_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V227_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V228_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V229_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V230_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V231_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V232_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V233_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V234_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V235_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V236_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V237_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V238_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V239_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V240_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V241_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V242_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V243_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V244_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V245_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V246_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V247_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V248_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V249_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V250_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V251_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V252_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V253_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V254_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V255_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V256_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V257_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V258_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V259_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V260_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V261_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V262_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V263_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V264_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V265_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V266_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V267_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V268_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V269_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V270_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V271_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V272_missing_flag                        ,  chisquared=10023.65463,   p-value=0.00000
V273_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V274_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V275_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V276_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V277_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V278_missing_flag                        ,  chisquared=9230.05652,   p-value=0.00000
V279_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V280_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V281_missing_flag                        ,  chisquared=0.02818,   p-value=0.86669
V282_missing_flag                        ,  chisquared=0.02818,   p-value=0.86669
V283_missing_flag                        ,  chisquared=0.02818,   p-value=0.86669
V284_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V285_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V286_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V287_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V288_missing_flag                        ,  chisquared=0.02818,   p-value=0.86669
V289_missing_flag                        ,  chisquared=0.02818,   p-value=0.86669
V290_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V291_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V292_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V293_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V294_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V295_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V296_missing_flag                        ,  chisquared=0.02818,   p-value=0.86669
V297_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V298_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V299_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V300_missing_flag                        ,  chisquared=0.02818,   p-value=0.86669
V301_missing_flag                        ,  chisquared=0.02818,   p-value=0.86669
V302_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V303_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V304_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V305_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V306_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V307_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V308_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V309_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V310_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V311_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V312_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V313_missing_flag                        ,  chisquared=0.02818,   p-value=0.86669
V314_missing_flag                        ,  chisquared=0.02818,   p-value=0.86669
V315_missing_flag                        ,  chisquared=0.02818,   p-value=0.86669
V316_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V317_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V318_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V319_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V320_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V321_missing_flag                        ,  chisquared=2.87936,   p-value=0.08972
V322_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V323_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V324_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V325_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V326_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V327_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V328_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V329_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V330_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V331_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V332_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V333_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V334_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V335_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V336_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V337_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V338_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
V339_missing_flag                        ,  chisquared=270.16694,   p-value=0.00000
id_01_missing_flag                       ,  chisquared=429.43113,   p-value=0.00000
id_02_missing_flag                       ,  chisquared=418.79840,   p-value=0.00000
id_03_missing_flag                       ,  chisquared=155.60340,   p-value=0.00000
id_04_missing_flag                       ,  chisquared=155.60340,   p-value=0.00000
id_05_missing_flag                       ,  chisquared=410.62370,   p-value=0.00000
id_06_missing_flag                       ,  chisquared=410.62370,   p-value=0.00000
id_07_missing_flag                       ,  chisquared=11.67033,   p-value=0.00064
id_08_missing_flag                       ,  chisquared=11.67033,   p-value=0.00064
id_09_missing_flag                       ,  chisquared=186.68294,   p-value=0.00000
id_10_missing_flag                       ,  chisquared=186.68294,   p-value=0.00000
id_11_missing_flag                       ,  chisquared=419.74461,   p-value=0.00000
id_12_missing_flag                       ,  chisquared=429.43113,   p-value=0.00000
id_13_missing_flag                       ,  chisquared=372.26945,   p-value=0.00000
id_14_missing_flag                       ,  chisquared=186.57626,   p-value=0.00000
id_15_missing_flag                       ,  chisquared=419.89703,   p-value=0.00000
id_16_missing_flag                       ,  chisquared=365.88234,   p-value=0.00000
id_17_missing_flag                       ,  chisquared=421.50054,   p-value=0.00000
id_18_missing_flag                       ,  chisquared=108.11199,   p-value=0.00000
id_19_missing_flag                       ,  chisquared=421.07000,   p-value=0.00000
id_20_missing_flag                       ,  chisquared=419.82460,   p-value=0.00000
id_21_missing_flag                       ,  chisquared=11.73422,   p-value=0.00061
id_22_missing_flag                       ,  chisquared=11.89450,   p-value=0.00056
id_23_missing_flag                       ,  chisquared=11.89450,   p-value=0.00056
id_24_missing_flag                       ,  chisquared=9.86160,   p-value=0.00169
id_25_missing_flag                       ,  chisquared=12.35494,   p-value=0.00044
id_26_missing_flag                       ,  chisquared=11.79823,   p-value=0.00059
id_27_missing_flag                       ,  chisquared=11.89450,   p-value=0.00056
id_28_missing_flag                       ,  chisquared=419.74461,   p-value=0.00000
id_29_missing_flag                       ,  chisquared=419.74461,   p-value=0.00000
id_30_missing_flag                       ,  chisquared=175.29045,   p-value=0.00000
id_31_missing_flag                       ,  chisquared=417.52693,   p-value=0.00000
id_32_missing_flag                       ,  chisquared=175.65824,   p-value=0.00000
id_33_missing_flag                       ,  chisquared=171.60347,   p-value=0.00000
id_34_missing_flag                       ,  chisquared=180.63611,   p-value=0.00000
id_35_missing_flag                       ,  chisquared=419.89703,   p-value=0.00000
id_36_missing_flag                       ,  chisquared=419.89703,   p-value=0.00000
id_37_missing_flag                       ,  chisquared=419.89703,   p-value=0.00000
id_38_missing_flag                       ,  chisquared=419.89703,   p-value=0.00000
DeviceType_missing_flag                  ,  chisquared=417.45054,   p-value=0.00000
DeviceInfo_missing_flag                  ,  chisquared=317.74598,   p-value=0.00000

# Significant variables
# print(significant_categorical_variables)

Calculate odds¶

Chi-Square test tells if the entire variable is useful or not.

ctab = pd.crosstab(df['ProductCD'], df['isFraud'].astype('category'))
ctab

Odds¶

ctab.columns = ctab.columns.add_categories('odds')
ctab['odds'] = ctab[1]/ctab[0]
ctab

Odds Ratio¶

ctab.columns = ctab.columns.add_categories('odds_ratio')
ctab['odds_ratio'] = ctab['odds'] / (ctab[1].sum()/ctab[0].sum())
ctab

Highers odds ratio implies more chance of fraud in that category.

Farther away it is from 1.0 (both directions) more important the variable is.

8. ANOVA Test¶

from scipy.stats import f_oneway

# significance value
alpha = 0.05

significant_numerical_variables = []
for col in num_columns[2:]:
    # Determine whether to reject or keep your null hypothesis
    if df.loc[:, col].nunique() > 50:
        F, p = f_oneway(df[df.isFraud == 1][col].dropna(),
                    df[df.isFraud == 0][col].dropna())
        print(col.ljust(40), ',   F-statistic=%.5f, p=%.5f' % (F, p), df.loc[:, col].nunique())
        if p <= alpha:
            significant_numerical_variables.append(col)

TransactionAmt                           ,   F-statistic=75.67718, p=0.00000 8195
card1                                    ,   F-statistic=109.88932, p=0.00000 13553
card2                                    ,   F-statistic=6.67558, p=0.00977 500
card3                                    ,   F-statistic=14336.21578, p=0.00000 114
card5                                    ,   F-statistic=661.82659, p=0.00000 119
addr1                                    ,   F-statistic=16.43435, p=0.00005 332
addr2                                    ,   F-statistic=485.06660, p=0.00000 74
dist1                                    ,   F-statistic=110.42966, p=0.00000 2412
C1                                       ,   F-statistic=552.37693, p=0.00000 1495
C2                                       ,   F-statistic=819.61843, p=0.00000 1167
C4                                       ,   F-statistic=545.60992, p=0.00000 1223
C5                                       ,   F-statistic=559.06343, p=0.00000 319
C6                                       ,   F-statistic=258.28478, p=0.00000 1291
C7                                       ,   F-statistic=468.69042, p=0.00000 1069
C8                                       ,   F-statistic=610.57329, p=0.00000 1130
C9                                       ,   F-statistic=594.15081, p=0.00000 205
C10                                      ,   F-statistic=476.56495, p=0.00000 1122
C11                                      ,   F-statistic=446.39851, p=0.00000 1343
C12                                      ,   F-statistic=601.74057, p=0.00000 1066
C13                                      ,   F-statistic=73.37132, p=0.00000 1464
C14                                      ,   F-statistic=37.04987, p=0.00000 1108
D1                                       ,   F-statistic=2672.56121, p=0.00000 641
D2                                       ,   F-statistic=2179.12220, p=0.00000 641
D3                                       ,   F-statistic=703.03864, p=0.00000 649
D4                                       ,   F-statistic=1913.51964, p=0.00000 808
D5                                       ,   F-statistic=1177.67893, p=0.00000 688
D6                                       ,   F-statistic=240.53556, p=0.00000 829
D8                                       ,   F-statistic=1555.93433, p=0.00000 5367
D10                                      ,   F-statistic=2681.29292, p=0.00000 818
D11                                      ,   F-statistic=634.20722, p=0.00000 676
D12                                      ,   F-statistic=53.96035, p=0.00000 635
D13                                      ,   F-statistic=219.58119, p=0.00000 577
D14                                      ,   F-statistic=4.66673, p=0.03076 802
D15                                      ,   F-statistic=3031.40845, p=0.00000 859
V37                                      ,   F-statistic=13626.03770, p=0.00000 55
V38                                      ,   F-statistic=17383.75607, p=0.00000 55
V56                                      ,   F-statistic=1940.19590, p=0.00000 52
V95                                      ,   F-statistic=10.01732, p=0.00155 881
V96                                      ,   F-statistic=17.75408, p=0.00003 1410
V97                                      ,   F-statistic=11.90221, p=0.00056 976
V99                                      ,   F-statistic=102.52878, p=0.00000 89
V101                                     ,   F-statistic=13.10561, p=0.00029 870
V102                                     ,   F-statistic=13.87026, p=0.00020 1285
V103                                     ,   F-statistic=15.47712, p=0.00008 928
V105                                     ,   F-statistic=6.15509, p=0.01310 100
V106                                     ,   F-statistic=2.58112, p=0.10815 56
V126                                     ,   F-statistic=7.41985, p=0.00645 10299
V127                                     ,   F-statistic=4.07834, p=0.04344 24414
V128                                     ,   F-statistic=2.19911, p=0.13809 14507
V129                                     ,   F-statistic=95.66082, p=0.00000 1608
V130                                     ,   F-statistic=5.58774, p=0.01809 5511
V131                                     ,   F-statistic=368.35409, p=0.00000 3097
V132                                     ,   F-statistic=10.72701, p=0.00106 6560
V133                                     ,   F-statistic=5.91363, p=0.01502 9949
V134                                     ,   F-statistic=7.27959, p=0.00697 8178
V135                                     ,   F-statistic=0.04505, p=0.83192 3724
V136                                     ,   F-statistic=0.00105, p=0.97410 4852
V137                                     ,   F-statistic=0.00778, p=0.92972 4252
V143                                     ,   F-statistic=72.67628, p=0.00000 870
V144                                     ,   F-statistic=221.93537, p=0.00000 63
V145                                     ,   F-statistic=315.85937, p=0.00000 260
V150                                     ,   F-statistic=336.66794, p=0.00000 1344
V151                                     ,   F-statistic=304.74317, p=0.00000 56
V159                                     ,   F-statistic=323.02425, p=0.00000 2492
V160                                     ,   F-statistic=319.70138, p=0.00000 9621
V161                                     ,   F-statistic=130.04611, p=0.00000 79
V162                                     ,   F-statistic=233.47479, p=0.00000 185
V163                                     ,   F-statistic=177.81176, p=0.00000 106
V164                                     ,   F-statistic=61.77344, p=0.00000 1978
V165                                     ,   F-statistic=217.06800, p=0.00000 2547
V166                                     ,   F-statistic=71.06800, p=0.00000 987
V167                                     ,   F-statistic=23.07236, p=0.00000 873
V168                                     ,   F-statistic=35.84863, p=0.00000 965
V171                                     ,   F-statistic=6876.38875, p=0.00000 62
V177                                     ,   F-statistic=26.27124, p=0.00000 862
V178                                     ,   F-statistic=38.22487, p=0.00000 1236
V179                                     ,   F-statistic=35.32729, p=0.00000 921
V180                                     ,   F-statistic=12.19456, p=0.00048 84
V182                                     ,   F-statistic=14.26650, p=0.00016 84
V187                                     ,   F-statistic=82.01146, p=0.00000 215
V201                                     ,   F-statistic=16856.02855, p=0.00000 56
V202                                     ,   F-statistic=45.39293, p=0.00000 10970
V203                                     ,   F-statistic=55.54978, p=0.00000 14951
V204                                     ,   F-statistic=60.13305, p=0.00000 12858
V205                                     ,   F-statistic=13.67504, p=0.00022 1953
V206                                     ,   F-statistic=4.52806, p=0.03335 1581
V207                                     ,   F-statistic=21.53123, p=0.00000 2705
V208                                     ,   F-statistic=11.70640, p=0.00062 2093
V209                                     ,   F-statistic=49.58723, p=0.00000 2674
V210                                     ,   F-statistic=0.05960, p=0.80713 2262
V211                                     ,   F-statistic=48.17990, p=0.00000 7624
V212                                     ,   F-statistic=62.41706, p=0.00000 8868
V213                                     ,   F-statistic=59.28563, p=0.00000 8317
V214                                     ,   F-statistic=3.28560, p=0.06989 2282
V215                                     ,   F-statistic=3.21830, p=0.07282 2747
V216                                     ,   F-statistic=0.05228, p=0.81914 2532
V217                                     ,   F-statistic=273.98881, p=0.00000 304
V218                                     ,   F-statistic=295.75410, p=0.00000 401
V219                                     ,   F-statistic=277.07672, p=0.00000 379
V221                                     ,   F-statistic=2528.82003, p=0.00000 77
V222                                     ,   F-statistic=4026.25520, p=0.00000 76
V224                                     ,   F-statistic=0.17276, p=0.67767 79
V226                                     ,   F-statistic=0.67393, p=0.41169 81
V228                                     ,   F-statistic=10162.77697, p=0.00000 55
V229                                     ,   F-statistic=2503.20448, p=0.00000 91
V230                                     ,   F-statistic=7401.92756, p=0.00000 66
V231                                     ,   F-statistic=222.73487, p=0.00000 294
V232                                     ,   F-statistic=390.01638, p=0.00000 338
V233                                     ,   F-statistic=279.30291, p=0.00000 333
V234                                     ,   F-statistic=53.01029, p=0.00000 122
V245                                     ,   F-statistic=1189.20909, p=0.00000 58
V253                                     ,   F-statistic=50.39006, p=0.00000 66
V258                                     ,   F-statistic=12632.06226, p=0.00000 67
V259                                     ,   F-statistic=3328.22023, p=0.00000 68
V263                                     ,   F-statistic=28.30710, p=0.00000 10422
V264                                     ,   F-statistic=11.42220, p=0.00073 13358
V265                                     ,   F-statistic=26.39408, p=0.00000 11757
V266                                     ,   F-statistic=3.60378, p=0.05765 1871
V267                                     ,   F-statistic=7.12289, p=0.00761 2884
V268                                     ,   F-statistic=3.47409, p=0.06234 2286
V269                                     ,   F-statistic=4.54627, p=0.03299 151
V270                                     ,   F-statistic=34.78841, p=0.00000 1972
V271                                     ,   F-statistic=105.06963, p=0.00000 2286
V272                                     ,   F-statistic=91.15867, p=0.00000 2082
V273                                     ,   F-statistic=32.05464, p=0.00000 4689
V274                                     ,   F-statistic=31.41484, p=0.00000 8315
V275                                     ,   F-statistic=38.23804, p=0.00000 4965
V276                                     ,   F-statistic=12.22436, p=0.00047 2263
V277                                     ,   F-statistic=15.03960, p=0.00011 2540
V278                                     ,   F-statistic=13.14749, p=0.00029 2398
V279                                     ,   F-statistic=8.30948, p=0.00394 881
V280                                     ,   F-statistic=0.36129, p=0.54779 975
V283                                     ,   F-statistic=7585.01985, p=0.00000 62
V285                                     ,   F-statistic=49.20336, p=0.00000 96
V290                                     ,   F-statistic=953.42763, p=0.00000 58
V291                                     ,   F-statistic=246.83555, p=0.00000 219
V292                                     ,   F-statistic=423.76303, p=0.00000 173
V293                                     ,   F-statistic=11.96267, p=0.00054 870
V294                                     ,   F-statistic=10.61896, p=0.00112 1286
V295                                     ,   F-statistic=1.66959, p=0.19631 928
V296                                     ,   F-statistic=10.20032, p=0.00140 94
V298                                     ,   F-statistic=0.98916, p=0.31995 94
V306                                     ,   F-statistic=2.04992, p=0.15222 16210
V307                                     ,   F-statistic=20.89358, p=0.00000 37367
V308                                     ,   F-statistic=6.25009, p=0.01242 23064
V309                                     ,   F-statistic=242.35559, p=0.00000 3239
V310                                     ,   F-statistic=72.36754, p=0.00000 7759
V311                                     ,   F-statistic=0.99820, p=0.31775 2526
V312                                     ,   F-statistic=835.01886, p=0.00000 5143
V313                                     ,   F-statistic=1016.37502, p=0.00000 3915
V314                                     ,   F-statistic=876.33173, p=0.00000 5974
V315                                     ,   F-statistic=1377.84553, p=0.00000 4540
V316                                     ,   F-statistic=5.17498, p=0.02291 9814
V317                                     ,   F-statistic=14.82121, p=0.00012 15184
V318                                     ,   F-statistic=0.58715, p=0.44352 12309
V319                                     ,   F-statistic=0.00218, p=0.96276 4799
V320                                     ,   F-statistic=14.53340, p=0.00014 6439
V321                                     ,   F-statistic=1.66098, p=0.19747 5560
V322                                     ,   F-statistic=38.22813, p=0.00000 881
V323                                     ,   F-statistic=44.84077, p=0.00000 1411
V324                                     ,   F-statistic=47.48458, p=0.00000 976
V329                                     ,   F-statistic=43.96148, p=0.00000 100
V330                                     ,   F-statistic=36.90161, p=0.00000 56
V331                                     ,   F-statistic=39.81114, p=0.00000 1758
V332                                     ,   F-statistic=45.37959, p=0.00000 2453
V333                                     ,   F-statistic=47.99171, p=0.00000 1971
V334                                     ,   F-statistic=0.01673, p=0.89708 143
V335                                     ,   F-statistic=2.44629, p=0.11781 669
V336                                     ,   F-statistic=0.47508, p=0.49066 355
V337                                     ,   F-statistic=2.67789, p=0.10175 254
V338                                     ,   F-statistic=30.86541, p=0.00000 380
V339                                     ,   F-statistic=17.71005, p=0.00003 334
id_01                                    ,   F-statistic=0.21069, p=0.64623 77
id_02                                    ,   F-statistic=5.40582, p=0.02007 115655
id_05                                    ,   F-statistic=0.62754, p=0.42826 93
id_06                                    ,   F-statistic=0.32488, p=0.56869 101
id_10                                    ,   F-statistic=0.15965, p=0.68948 62
id_11                                    ,   F-statistic=0.74750, p=0.38727 146
id_13                                    ,   F-statistic=1.07632, p=0.29952 54
id_17                                    ,   F-statistic=1.92097, p=0.16575 104
id_19                                    ,   F-statistic=0.09828, p=0.75391 522
id_20                                    ,   F-statistic=0.34673, p=0.55597 394
Date                                     ,   F-statistic=101.40691, p=0.00000 573349

# Significant variables
# significant_numerical_variables

EDA Inferences:

The target class in imbalanced
Only 3.5% transactions are fraud in terms of count and 3.87% in terms of transaction amount
TransactionAmt is right skewed so log transform needs to be used to make it normally distributed
Fraud Transaction rate is maximum for Product Category C and minimum for Product Category W
97% of transactions are from Mastercard(32%) and Visa(65%
Fraud transaction rate is highest for discover cards(~8%) against ~3.5% of Mastercard and Visa and 2.87% in American Express
Almost all the transactions are from Credit and Debit cards.
Debit card transactions are almost 3 times as compared to credit card transactions.
Fraud transaction rate is high for Credit cards as compared to Debit cards.
Fraud transaction rate for Microsoft is high as compared to Google and Yahoo mail # p emaildomain
Fraud transaction rate (amount) for Google is high as comapred to Microsoft and Yahoo mail #p emaildomain
There isn't any information about R_emaildomain for Majority of transactions (76.75% count , 85.62% amount) #r emaildomain
Fraud transaction rate for Google is high as compared to Yahoo, anaonymous.com and Microsoft #r emaildomain
Surprisingly fraud transaction rate is high on the days when number of transactions are less
Day 29,30 and 31 are having less transactions, looks like people are broke at the month end
Surprisingly fraud transaction rate is high on the days when number of transactions and transaction amounts are less. Day 0 and 6
Day 0 and 6 have less transactions, these might be weekend days
Transactions start decreasing mid night but the fraud rate starts increasing
Transactions from 3 AM to 12 PM needs to monitored very closely

Mini Challenge¶

Analyze the columns having majority of missing values. Create a new column which has a flag whether the value is missing or not, later on try to find out some pattern of missing data with the target class

7. Feature Engineering ¶

Feature engineering is the process of using domain and statistical knowledge to extract features from raw data via data mining techniques.

These features often help to improve the performance of machine learning models.

The goal of this section is to:

Engineer domain specific features
Dimensionality reduction
Encode the categorical features

df.head()

Domain Specific Features¶

You need to engineer the domain specific features. This might boost up the predictive power. This often gives better performing models

Domain knowledge is one of the key pillars of data science. So always understand the domain before attempting the problem.

# Transaction amount minus mean of transaction 
df['Trans_min_mean'] = df['TransactionAmt'] - np.nanmean(df['TransactionAmt'],dtype="float64")
df['Trans_min_std']  = df['Trans_min_mean'] / np.nanstd(df['TransactionAmt'].astype("float64"),dtype="float64")

Replace value by the group's mean (or standard dev)

# Features for transaction amount and card 
df['TransactionAmt_to_mean_card1'] = df['TransactionAmt'] / df.groupby(['card1'])['TransactionAmt'].transform('mean')
df['TransactionAmt_to_mean_card4'] = df['TransactionAmt'] / df.groupby(['card4'])['TransactionAmt'].transform('mean')
df['TransactionAmt_to_std_card1']  = df['TransactionAmt'] / df.groupby(['card1'])['TransactionAmt'].transform('std')
df['TransactionAmt_to_std_card4']  = df['TransactionAmt'] / df.groupby(['card4'])['TransactionAmt'].transform('std')

# Log of transaction amount
df['TransactionAmt'] = np.log(df['TransactionAmt'])

df.head()

# Save train df to csv file 
# df.to_csv("Intermediate_Datasets/df_intermediate1.csv",index = False)

# Read train df
df = pd.read_csv("Intermediate_Datasets/df_intermediate1.csv")

8. Dimensionality Reduction - PCA ¶

When dealing with high dimensional data, it is often useful to reduce the dimensionality by projecting the data to a lower dimensional subspace which captures the “essence” of the data.

Dimensionality reduction, or dimension reduction, is the transformation of data from a high-dimensional space into a low-dimensional space so that the low-dimensional representation retains some meaningful properties of the original data, ideally close to its intrinsic dimension.

Principal component analysis is a technique for reducing the dimensionality of such datasets, increasing interpretability but at the same time minimizing information loss. It does so by creating new uncorrelated variables that successively maximize variance.

# initialize function to perform PCA
def perform_PCA(df, cols, n_components, prefix='PCA_', rand_seed=4):
    pca = PCA(n_components=n_components, random_state=rand_seed)
    principalComponents = pca.fit_transform(df[cols])
    principalDf = pd.DataFrame(principalComponents)
    df.drop(cols, axis=1, inplace=True)

    principalDf.rename(columns=lambda x: str(prefix)+str(x), inplace=True)
    df = pd.concat([df, principalDf], axis=1)
    return df

Create a list of all the columns on which PCA needs to performed

# Columns starting from V1 to V339
filter_col = df.columns[53:392]

Impute missing values in the mas_v columns, later use minmax_scale function to scale the values in these columns

from   sklearn.preprocessing import minmax_scale

# Fill na values and scale V columns
for col in filter_col:
    df[col] = df[col].fillna((df[col].min() - 2))
    df[col] = (minmax_scale(df[col], feature_range=(0,1)))

# Perform PCA    
df          = perform_PCA(df, filter_col, prefix='PCA_V_', n_components=30)

Reduce memory usage of df as lot of new features have been created

df = reduce_mem_usage(df)

Mem. usage decreased to 1138.99 Mb (21.4% reduction)

df.head()

# Plot first 2 PCA features and colour by target variable
plt.figure(figsize=(12, 8));
groups = df.groupby("isFraud")
for name, group in groups:
    plt.scatter(group["PCA_V_0"], group["PCA_V_1"], label=name)
plt.legend()
plt.show()

9. Feature Encoding ¶

Encoding is the process of converting data from one form to another. Most of the Machine learning algorithms can not handle categorical values unless we convert them to numerical values. Many algorithm’s performances vary based on how Categorical columns are encoded.

Frequency Encoding - It is a way to utilize the frequency of the categories as labels. In the cases where the frequency is related somewhat with the target variable, it helps the model to understand and assign the weight in direct and inverse proportion, depending on nature of the data.

Create a list of variables that needs to be encoded using frequency encoding. Let's note down the features which has more than 30 unique values, We would using frequency encoding for these features only

cat_columns = df.select_dtypes(include=['object']).columns
len(cat_columns)

30

binary_columns = [col for col in df.columns if df[col].nunique() == 2]
len(binary_columns)

432

num_columns = [col for col in df.columns if (col not in cat_columns) & (col not in binary_columns)]
len(num_columns)

92

cat_columns = cat_columns.to_list() + binary_columns

# Frequecny encoding variables
frequency_encoded_variables = []
for col in cat_columns:
    if df[col].nunique() > 30:
        print(col, df[col].nunique())
        frequency_encoded_variables.append(col)

id_33 260
DeviceInfo 1786
Date 573349

It's time to encode the variables using frequency encoding

# Frequecny enocde the variables
for variable in tqdm(frequency_encoded_variables):
    # group by frequency 
    fq = df.groupby(variable).size()/len(df)    
    # mapping values to dataframe 
    df.loc[:, "{}".format(variable)] = df[variable].map(fq)   
    cat_columns.remove(variable)

100%|████████████████████████████████████████████████████████████████████████████████████| 3/3 [00:01<00:00,  2.18it/s]

df.head()

Label encoding - Label Encoding refers to converting the labels into numeric form so as to convert it into the machine-readable form. Machine learning algorithms can then decide in a better way on how those labels must be operated. It is an important pre-processing step for the structured dataset in supervised learning.

It is a popular encoding technique for handling categorical variables. In this technique, each label is assigned a unique integer based on alphabetical ordering.

# Label encode the variables
for col in cat_columns:
    lbl        = LabelEncoder()
    lbl.fit(list(df[col].values))
    df[col] = lbl.transform(list(df[col].values))

Let's reduce the memory usage as lot of new columns has been added to the data frame

# Reduce memory usage
df = reduce_mem_usage(df)

Mem. usage decreased to 361.00 Mb (82.7% reduction)

Tip : Save the train df, and clean all memory

# Save train df to csv file 
df.to_csv("Intermediate_Datasets/df_intermediate2.csv", index = False)

10. Data Preprocessing for Model Building ¶

The goal of this section is to:

Clean up columns
Create X and y
Split the dataset in training and test sets

# Read train df
df = pd.read_csv("Intermediate_Datasets/df_intermediate2.csv")
# df = df.sample(10000, random_state=0)

df.loc[:, 'isFraud'].value_counts()

0    569877
1     20663
Name: isFraud, dtype: int64

Drop the columns which may not be useful for model building

df = df.drop(['TransactionID','TransactionDT','Date'], axis=1)

Separate the x variables and y variables

# Split the y variable series and x variables dataset
X = df.drop(['isFraud'],axis=1)
y = df.isFraud.astype(bool)

# Delete train df
del df

# Collect garbage
gc.collect()

0

Split the dataset into train set and test set. Train set will be used to train the model. Test set will be used to check the performance of model

# Split the dataset into the training set and test set
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.3, random_state = 0)

# Head of X_train
X_train.head()

11. Model Building ¶

Finally, model building starts here.

The goal of this section is to:

Build ML models
Evaluate the performance

12. XGBoost Classifier¶

XGBoost is an optimized distributed gradient boosting model designed to be highly efficient, flexible and portable. It implements machine learning algorithms under the Gradient Boosting framework. XGBoost provides a parallel tree boosting (also known as GBDT, GBM) that solve many data science problems in a fast and accurate way.

%%time
# Define the model
xgb = XGBClassifier(nthread = -1, random_state=0)

# Train the model
xgb.fit(X_train, y_train)

xgb

Wall time: 10min 13s

XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,
              colsample_bynode=1, colsample_bytree=1, gamma=0, gpu_id=-1,
              importance_type='gain', interaction_constraints='',
              learning_rate=0.300000012, max_delta_step=0, max_depth=6,
              min_child_weight=1, missing=nan, monotone_constraints='()',
              n_estimators=100, n_jobs=-1, nthread=-1, num_parallel_tree=1,
              random_state=0, reg_alpha=0, reg_lambda=1, scale_pos_weight=1,
              subsample=1, tree_method='exact', validate_parameters=1,
              verbosity=None)

Let's use the model to get predictions on test dataset. We would be looking at the predicted class and predicted probability both in order to evaluate the performance of the model

# Prediction
y_pred_xgb = xgb.predict(X_test)
y_prob_pred_xgb = xgb.predict_proba(X_test)
y_prob_pred_xgb = [x[1] for x in y_prob_pred_xgb]
print("Y predicted : ",y_pred_xgb)
print("Y probability predicted : ",y_prob_pred_xgb[:5])

Y predicted :  [False False False ... False False False]
Y probability predicted :  [0.00088362244, 0.0128200315, 0.003954415, 0.008103509, 0.0021388768]

13. Evaluation Metrics ¶

Accuracy Score
Confusion Matrix
Classification Report
AUC Score
Concordance Index
ROC curve
PR curve

Concordance

from bisect import bisect_left, bisect_right

def concordance(actuals, preds):
    ones_preds  = [p for a,p in zip(actuals, preds) if a == 1]
    zeros_preds = [p for a,p in zip(actuals, preds) if a == 0]
    n_ones      = len([x for x in actuals if x == 1])
    n_total_pairs =  float(n_ones) * float(len(actuals) - n_ones)
    # print("Total Pairs: ", n_total_pairs)

    zeros_sorted = sorted(zeros_preds)

    conc = 0; disc = 0; ties = 0;
    for i, one_pred in enumerate(ones_preds):
        cur_conc = bisect_left(zeros_sorted, one_pred)
        cur_ties = bisect_right(zeros_sorted, one_pred) - cur_conc
        conc += cur_conc
        ties += cur_ties
        disc += float(len(zeros_sorted)) - cur_ties - cur_conc

    concordance = conc/n_total_pairs
    discordance = disc/n_total_pairs
    ties_perc = ties/n_total_pairs
    return concordance

All evaluation metrics

def compute_evaluation_metric(model, x_test, y_actual, y_predicted, y_predicted_prob):
    print("\n Accuracy Score : ",accuracy_score(y_actual,y_predicted))
    print("\n AUC Score : ", roc_auc_score(y_actual, y_predicted_prob))
    print("\n Confusion Matrix : \n",confusion_matrix(y_actual, y_predicted))
    print("\n Classification Report : \n",classification_report(y_actual, y_predicted))
    print("\n Concordance Index : ", concordance(y_actual, y_predicted_prob))

    print("\n ROC curve : \n")
    plot_roc_curve(model, x_test, y_actual)
    plt.show() 

    print("\n PR curve : \n")
    plot_precision_recall_curve(model, x_test, y_actual)
    plt.show()

concordance(y_test.values, y_prob_pred_xgb)

0.9348031547662842

# Compute Evaluation Metric
compute_evaluation_metric(xgb, X_test, y_test, y_pred_xgb, y_prob_pred_xgb)

 Accuracy Score :  0.9797247716778993

 AUC Score :  0.9348031681385632

 Confusion Matrix : 
 [[170693    348]
 [  3244   2877]]

 Classification Report : 
               precision    recall  f1-score   support

       False       0.98      1.00      0.99    171041
        True       0.89      0.47      0.62      6121

    accuracy                           0.98    177162
   macro avg       0.94      0.73      0.80    177162
weighted avg       0.98      0.98      0.98    177162


 Concordance Index :  0.9348031547662842

 ROC curve :

 PR curve :

14. Capture Rates and Calibration Curve ¶

Divide the data in 10 equal bins as per predicted probability scores. Then, compute the percentage of the total target class 1 captured in every bin.

Ideally the proportion should be decreasing as we go down ever bin. Let's check it out

Create validation set¶

# Create Validation set
validation_df = {'y_test' : y_test, 'y_pred' : y_pred_xgb, 'y_pred_prob' : y_prob_pred_xgb}
validation_df = pd.DataFrame(data = validation_df)

# Add binning column to the dataframe
validation_df['bin_y_pred_prob'] = pd.qcut(validation_df['y_pred_prob'], q=10)
validation_df.head()

# Change x label
x_label = []
for i in range(len(validation_df['bin_y_pred_prob'].cat.categories[::-1].astype('str'))):
    x_label.append("Bin" + str(i + 1)+ "(" + validation_df['bin_y_pred_prob'].cat.categories[::-1].astype('str')[i] + ")")

Capture Rates Plot¶

# Plot Distribution of predicted probabilities for every bin
plt.figure(figsize=(12, 8));
sns.stripplot(validation_df.bin_y_pred_prob, validation_df.y_pred_prob, jitter = 0.15, hue = validation_df.y_test, order = validation_df['bin_y_pred_prob'].cat.categories[::-1])
plt.title("Distribution of predicted probabilities for every bin", fontsize=18)
plt.xlabel("Predicted Probability Bins", fontsize=14);
plt.ylabel("Predicted Probability", fontsize=14);
plt.xticks(np.arange(10), x_label, rotation=45);
plt.show()

Gains Table¶

# Aggregate the data
gains_df             = validation_df.groupby(["bin_y_pred_prob","y_test"]).agg({'y_test': ['count']})
gains_df.columns     = gains_df.columns.map(''.join)
gains_df['prob_bin'] = gains_df.index.get_level_values(0)
gains_df['y_test']   = gains_df.index.get_level_values(1)
gains_df.reset_index(drop = True, inplace = True)
gains_df

# Get infection rate and percentage infections
gains_table = gains_df.pivot(index='prob_bin', columns='y_test', values='y_testcount')
gains_table['prob_bin'] = gains_table.index
gains_table = gains_table.iloc[::-1]
gains_table['prob_bin'] = x_label
gains_table.reset_index(drop = True, inplace = True)
gains_table = gains_table[['prob_bin', 0, 1]]
gains_table.columns = ['prob_bin', "not_fraud", "fraud"]
gains_table['perc_fraud'] = gains_table['fraud']/gains_table['fraud'].sum()
gains_table['perc_not_fraud'] = gains_table['not_fraud']/gains_table['not_fraud'].sum()
gains_table['cum_perc_fraud'] = 100*(gains_table.fraud.cumsum() / gains_table.fraud.sum()) 
gains_table['cum_perc_not_fraud'] = 100*(gains_table.not_fraud.cumsum() / gains_table.not_fraud.sum()) 
gains_table


# Plot
plt.figure(figsize=(12, 8));
sns.set_style("white")
sns.pointplot(x = "prob_bin", y = "cum_perc_fraud", data = gains_table, legend = False, order=gains_table.prob_bin)
plt.xticks(rotation=45);
plt.ylabel("Fraud Rate", fontsize=14)
plt.xlabel("Prediction probability bin", fontsize=14)
plt.title("Fraud rate for every bin", fontsize=18)
plt.show()

Ideally the slope should be high initially and should decrease as we move further to the right. This is not really a good model.

# One big function.
def captures(y_test, y_pred, y_pred_prob):
    # Create Validation set
    validation_df = {'y_test' : y_test, 'y_pred' : y_pred, 'y_pred_prob' : y_pred_prob}
    validation_df = pd.DataFrame(data = validation_df)

    # Add binning column to the dataframe
    try:
        validation_df['bin_y_pred_prob'] = pd.qcut(validation_df['y_pred_prob'], q=10)
    except:
        validation_df['bin_y_pred_prob'] = pd.qcut(validation_df['y_pred_prob'], q=10, duplicates='drop')
    
    # Change x label and column names
    x_label = []
    for i in range(len(validation_df['bin_y_pred_prob'].cat.categories[::-1].astype('str'))):
        x_label.append("Bin" + str(i + 1)+ "(" + validation_df['bin_y_pred_prob'].cat.categories[::-1].astype('str')[i] + ")")
    
    # Plot Distribution of predicted probabilities for every bin
    plt.figure(figsize=(12, 8));
    sns.stripplot(validation_df.bin_y_pred_prob, validation_df.y_pred_prob, jitter = 0.15, hue = validation_df.y_test, order = validation_df['bin_y_pred_prob'].cat.categories[::-1])
    plt.title("Distribution of predicted probabilities for every bin", fontsize=18)
    plt.xlabel("Predicted Probability Bins", fontsize=14);
    plt.ylabel("Predicted Probability", fontsize=14);
    try:
        plt.xticks(np.arange(10), x_label, rotation=45);
    except:
        pass
    plt.show()
    
    # Aggregate the data
    gains_df             = validation_df.groupby(["bin_y_pred_prob","y_test"]).agg({'y_test': ['count']})
    gains_df.columns     = gains_df.columns.map(''.join)
    gains_df['prob_bin'] = gains_df.index.get_level_values(0)
    gains_df['y_test']   = gains_df.index.get_level_values(1)
    gains_df.reset_index(drop = True, inplace = True)
    gains_df

    # Get infection rate and percentage infections
    gains_table = gains_df.pivot(index='prob_bin', columns='y_test', values='y_testcount')
    gains_table['prob_bin'] = gains_table.index
    gains_table = gains_table.iloc[::-1]
    gains_table['prob_bin'] = x_label
    gains_table.reset_index(drop = True, inplace = True)
    gains_table = gains_table[['prob_bin', 0, 1]]
    gains_table.columns = ['prob_bin', "not_fraud", "fraud"]
    gains_table['perc_fraud'] = gains_table['fraud']/gains_table['fraud'].sum()
    gains_table['perc_not_fraud'] = gains_table['not_fraud']/gains_table['not_fraud'].sum()
    gains_table['cum_perc_fraud'] = 100*(gains_table.fraud.cumsum() / gains_table.fraud.sum()) 
    gains_table['cum_perc_not_fraud'] = 100*(gains_table.not_fraud.cumsum() / gains_table.not_fraud.sum()) 
    gains_table


    # Plot
    plt.figure(figsize=(12, 8));
    sns.set_style("white")
    sns.pointplot(x = "prob_bin", y = "cum_perc_fraud", data = gains_table, legend = False, order=gains_table.prob_bin)
    plt.xticks(rotation=45);
    plt.ylabel("Fraud Rate", fontsize=14)
    plt.xlabel("Prediction probability bin", fontsize=14)
    plt.title("Fraud rate for every bin", fontsize=18)
    plt.show()
    return gains_table

# Gains Table and Capture rates plot
captures(y_test, y_pred_xgb, y_prob_pred_xgb)

Calibration Curve¶

from sklearn.calibration import calibration_curve
import matplotlib.pyplot as plt

def draw_calibration_curve(y_test, y_prob, n_bins=10):
    plt.figure(figsize=(7, 7), dpi=120)
    ax1 = plt.subplot2grid((3, 1), (0, 0), rowspan=2)
    ax2 = plt.subplot2grid((3, 1), (2, 0))
    ax1.plot([0, 1], [0, 1], "k:", label="Perfectly calibrated")


    fraction_of_positives, mean_predicted_value = calibration_curve(y_test, y_prob, n_bins=10)

    ax1.plot(mean_predicted_value, fraction_of_positives, "s-", label="%s" % ("Model", ))
    ax2.hist(y_prob, range=(0, 1), bins=10, label="Model", histtype="step", lw=2)

    # Labels
    ax1.set_ylabel("Fraction of positives")
    ax1.set_ylim([-0.05, 1.05])
    ax1.legend(loc="lower right")
    ax1.set_title('Calibration plots  (reliability curve)')

    ax2.set_xlabel("Mean predicted value")
    ax2.set_ylabel("Count")
    ax2.legend(loc="upper center", ncol=2)

    plt.tight_layout()
    plt.show()

Chart 1: X axis marks the prediction probability score. Y-axis marks the fraction of the positives.
Chart 2: X axis marks the mean predicted value. Y-axis represents the count of records.

draw_calibration_curve(y_test, y_prob_pred_xgb, n_bins=10)

Calibrate the model¶

Logistic regression

# Prediction
y_pred_xgb_test = xgb.predict(X_test)
y_prob_pred_xgb_test = xgb.predict_proba(X_test)[:, 1]

from sklearn.linear_model import LogisticRegression
X = np.array(y_prob_pred_xgb_test)
clf = LogisticRegression(random_state=0).fit(X.reshape(-1, 1), y_test)

y_prob_pred_calib = clf.predict_proba(X.reshape(-1, 1))[:, 1]
y_pred_calib      = clf.predict(X.reshape(-1, 1))

captures(y_test, y_pred_calib, y_prob_pred_calib)

draw_calibration_curve(y_test, y_prob_pred_calib, n_bins=10)

XGBoost with booster = dart¶

%%time
# Define the model
xgb = XGBClassifier(nthread=-1, random_state=0, booster="dart")

# Train the model
xgb.fit(X_train,y_train)

xgb

Wall time: 11min 20s

XGBClassifier(base_score=0.5, booster='dart', colsample_bylevel=1,
              colsample_bynode=1, colsample_bytree=1, gamma=0, gpu_id=-1,
              importance_type='gain', interaction_constraints='',
              learning_rate=0.300000012, max_delta_step=0, max_depth=6,
              min_child_weight=1, missing=nan, monotone_constraints='()',
              n_estimators=100, n_jobs=-1, nthread=-1, num_parallel_tree=1,
              random_state=0, reg_alpha=0, reg_lambda=1, scale_pos_weight=1,
              subsample=1, tree_method='exact', validate_parameters=1,
              verbosity=None)

Let's use the model to get predictions on test dataset. We would be looking at the predicted class and predicted probability both in order to evaluate the performance of the model

Prediction¶

# Prediction
y_pred_xgbdart      = xgb.predict(X_test)
y_prob_pred_xgbdart = xgb.predict_proba(X_test)[:, 1]
print("Y predicted : ", y_pred_xgbdart)
print("Y probability predicted : ", y_prob_pred_xgbdart[:5])

Y predicted :  [False False False ... False False False]
Y probability predicted :  [0.00088362 0.01282003 0.00395442 0.00810351 0.00213888]

Evaluation Metrices¶

Let's compute various evaluation metrices now

Accuracy Score
Confusion Matrix
Classification Report
AUC Score
Concodense Index
ROC curve
PR curve

# Compute Evaluation Metric
compute_evaluation_metric(xgb, X_test, y_test, y_pred_xgbdart, y_prob_pred_xgbdart)

 Accuracy Score :  0.9797247716778993

 AUC Score :  0.9348031681385632

 Confusion Matrix : 
 [[170693    348]
 [  3244   2877]]

 Classification Report : 
               precision    recall  f1-score   support

       False       0.98      1.00      0.99    171041
        True       0.89      0.47      0.62      6121

    accuracy                           0.98    177162
   macro avg       0.94      0.73      0.80    177162
weighted avg       0.98      0.98      0.98    177162


 Concordance Index :  0.9348031547662842

 ROC curve :

 PR curve :

# Gains Table and Capture rates plot
captures(y_test, y_pred_xgbdart, y_prob_pred_xgbdart)

draw_calibration_curve(y_test, y_prob_pred_xgbdart, n_bins=10)

Inferences:

Both of the boosters are giving the similar result, maybe because both are tree based
The accuracy score is 0.97, AUC and concordence scores are 0.88.
Recall and f1-score are very less for class True. That's because of class imbalance
ROC and PR curve also needs improvements

Let's look at LGBM

15. LightGBM ¶

LightGBM is a gradient boosting framework that uses tree based learning algorithms.

It is designed to be distributed and efficient with the following advantages:

Faster training speed and higher efficiency.
Lower memory usage.
Better accuracy.
Support of parallel and GPU learning.
Capable of handling large-scale data.

from lightgbm import LGBMClassifier

%%time
# Define the model
lgbc = LGBMClassifier(random_state=0, n_jobs = -1)

# Train the model
lgbc.fit(X_train,y_train)

lgbc

Wall time: 1min 16s

LGBMClassifier(random_state=0)

Let's use the model to get predictions on test dataset. We would be looking at the predicted class and predicted probability both in order to evaluate the performance of the model

# Prediction
y_pred_lgbc = lgbc.predict(X_test)
y_prob_pred_lgbc = lgbc.predict_proba(X_test)
y_prob_pred_lgbc = [x[1] for x in y_prob_pred_lgbc]
print("Y predicted : ",y_pred_lgbc)
print("Y probability predicted : ",y_prob_pred_lgbc[:5])

Y predicted :  [False False False ... False False False]
Y probability predicted :  [0.0030038692434659836, 0.0204190888379695, 0.017504758484994488, 0.0071321007747928164, 0.0018421571426377808]

Evaluation Metrices¶

Let's compute various evaluation metrices now

Accuracy Score
Confusion Matrix
Classification Report
AUC Score
Concodense Index
ROC curve
PR curve

# Compute Evaluation Metric
compute_evaluation_metric(lgbc, X_test, y_test, y_pred_lgbc, y_prob_pred_lgbc)

 Accuracy Score :  0.9772355245481537

 AUC Score :  0.9276554552960554

 Confusion Matrix : 
 [[170633    408]
 [  3625   2496]]

 Classification Report : 
               precision    recall  f1-score   support

       False       0.98      1.00      0.99    171041
        True       0.86      0.41      0.55      6121

    accuracy                           0.98    177162
   macro avg       0.92      0.70      0.77    177162
weighted avg       0.98      0.98      0.97    177162


 Concordance Index :  0.9276554118361485

 ROC curve :

 PR curve :

# Gains Table and Capture rates plot
captures(y_test, y_pred_lgbc, y_prob_pred_lgbc)

draw_calibration_curve(y_test, y_prob_pred_lgbc, n_bins=10)

Inferences:

With LGBM, Accuracy score is 97.7%. It's almost similar to XGBoost model
AUC score has imporved to 92.6 from 88.5
Recall and f-1 score have also improved, but it's still not upto the mark

16. Random Forest Classifier ¶

from sklearn.ensemble import RandomForestClassifier, ExtraTreesClassifier, GradientBoostingClassifier

X_train.head()

Impute Missing values. Since sklearn algos are not designed to handle missing values.

from sklearn.impute import KNNImputer, SimpleImputer

# replace inf
X_train = X_train.replace([np.inf, -np.inf], np.nan)
X_test = X_test.replace([np.inf, -np.inf], np.nan)

# Impute
imputer = SimpleImputer(missing_values=np.nan, strategy='mean')
# imputer = KNNImputer(n_neighbors=3)

X_train_imputed = imputer.fit_transform(X_train)
X_train_imputed = pd.DataFrame(X_train_imputed, columns=X_train.columns)
X_train_imputed.head()

Build and train the Classifier

%%time
# Define the model
rfc = RandomForestClassifier(random_state=0, n_jobs = -1)
# rfc = ExtraTreesClassifier(random_state=0, n_jobs = -1)
# rfc = AdaBoostClassifier(random_state=0)
# rfc = GradientBoostingClassifier(random_state=0)

# Train the model
rfc.fit(X_train_imputed, y_train)

rfc

Predicting on test data

# Impute X_Test before predicting
X_test_imputed = imputer.transform(X_test)

# Prediction
y_pred_rfc = rfc.predict(X_test_imputed)
y_prob_pred_rfc = rfc.predict_proba(X_test_imputed)[:, 1]

print("Y predicted : ",y_pred_rfc)
print("Y probability predicted : ",y_prob_pred_rfc[:5])

Y predicted :  [False False False ... False False False]
Y probability predicted :  [0.01302948 0.03184159 0.0321819  0.01066897 0.00670777]

Evaluation metrics

# Compute Evaluation Metric
compute_evaluation_metric(rfc, X_test_imputed, y_test, y_pred_rfc, y_prob_pred_rfc)

 Accuracy Score :  0.9736173671554849

 AUC Score :  0.8775531354407143

 Confusion Matrix : 
 [[170697    344]
 [  4330   1791]]

 Classification Report : 
               precision    recall  f1-score   support

       False       0.98      1.00      0.99    171041
        True       0.84      0.29      0.43      6121

    accuracy                           0.97    177162
   macro avg       0.91      0.65      0.71    177162
weighted avg       0.97      0.97      0.97    177162


 Concordance Index :  0.8775057827680287

 ROC curve :

 PR curve :

# Concordance
concordance(y_test.values, y_prob_pred_rfc)

0.8775057827680287

# Gains Table and Capture rates plot
captures(y_test, y_pred_rfc, y_prob_pred_rfc)

draw_calibration_curve(y_test, y_prob_pred_rfc, n_bins=10)

17. Handling Class Imbalance ¶

Handle Class Imbalance with Random Oversampler¶

Imbalanced classes are a common problem in machine learning classification where there are a disproportionate ratio of observations in each class.

Most machine learning algorithms work best when the number of samples in each class are about equal. This is because most algorithms are designed to maximize accuracy and reduce error.

Upsample

# random over sampler
ros = RandomOverSampler()
X_train_ros, y_train_ros = ros.fit_sample(X_train_imputed, y_train)
y_train_ros.value_counts()

True     398836
False    398836
Name: isFraud, dtype: int64

%%time
# Define the model
lgbc_ros = LGBMClassifier(random_state=0)

# Train the model
lgbc_ros.fit(X_train_ros,y_train_ros)

lgbc_ros

Wall time: 3min 20s

LGBMClassifier(random_state=0)

Let's use the model to get predictions on test dataset. We would be looking at the predicted class and predicted probability both in order to evaluate the performance of the model

# Prediction on the original test dataset
y_pred_lgbcros = lgbc_ros.predict(X_test_imputed)
y_prob_pred_lgbcros = lgbc_ros.predict_proba(X_test_imputed)[:, 1]

print("Y predicted : ",y_pred_lgbcros)
print("Y probability predicted : ",y_prob_pred_lgbcros[:5])

Y predicted :  [False False False ... False False  True]
Y probability predicted :  [0.04800962 0.30784135 0.30226978 0.104903   0.03731646]

Evaluation Metrics¶

Let's compute various evaluation metrices now

Accuracy Score
Confusion Matrix
Classification Report
AUC Score
Concodense Index
ROC curve
PR curve

# Compute Evaluation Metric
compute_evaluation_metric(lgbc_ros, X_test_imputed, y_test, y_pred_lgbcros, y_prob_pred_lgbcros)

 Accuracy Score :  0.8851503144015083

 AUC Score :  0.9252961635759673

 Confusion Matrix : 
 [[151857  19184]
 [  1163   4958]]

 Classification Report : 
               precision    recall  f1-score   support

       False       0.99      0.89      0.94    171041
        True       0.21      0.81      0.33      6121

    accuracy                           0.89    177162
   macro avg       0.60      0.85      0.63    177162
weighted avg       0.97      0.89      0.92    177162


 Concordance Index :  0.9252961416072233

 ROC curve :

 PR curve :

# Gains Table and Capture rates plot
captures(y_test, y_pred_lgbcros, y_prob_pred_lgbcros)

draw_calibration_curve(y_test, y_prob_pred_lgbcros, n_bins=10)

Inferences:

After balancing the class, accuracy score is 0.88 and AUC score is 92.5%
Accuracy has decreased as compared to the previos model, but AUC has improved
Additionally the recall has improved significantly at the cost of precision.

18. Cost Sensitive Learning with Class weights ¶

The 'balanced' mode uses the values of y to automatically adjust weights inversely proportional to class frequencies in the input data as n_samples / (n_classes * np.bincount(y))

%%time
# Define the model
lgbc_bal = LGBMClassifier(random_state=0, class_weight='balanced')

# Train the model
lgbc_bal.fit(X_train_imputed, y_train)

lgbc_bal

Wall time: 1min 30s

LGBMClassifier(class_weight='balanced', random_state=0)

# Prediction
y_pred_lgbcbal = lgbc_bal.predict(X_test)
y_prob_pred_lgbcbal = lgbc_bal.predict_proba(X_test)[:, 1]

print("Y predicted : ",y_pred_lgbcbal)
print("Y probability predicted : ",y_prob_pred_lgbcbal[:5])

Y predicted :  [False False False ... False False  True]
Y probability predicted :  [0.17759182 0.43730851 0.42168771 0.41189862 0.17474149]

# Compute Evaluation Metric
compute_evaluation_metric(lgbc_bal, X_test, y_test, y_pred_lgbcbal, y_prob_pred_lgbcbal)

 Accuracy Score :  0.7243144692428399

 AUC Score :  0.9084094113408069

 Confusion Matrix : 
 [[122883  48158]
 [   683   5438]]

 Classification Report : 
               precision    recall  f1-score   support

       False       0.99      0.72      0.83    171041
        True       0.10      0.89      0.18      6121

    accuracy                           0.72    177162
   macro avg       0.55      0.80      0.51    177162
weighted avg       0.96      0.72      0.81    177162


 Concordance Index :  0.9084093946254581

 ROC curve :

 PR curve :

# Gains Table and Capture rates plot
captures(y_test, y_pred_lgbcbal, y_prob_pred_lgbcbal)

draw_calibration_curve(y_test, y_prob_pred_lgbcbal, n_bins=10)

19. Model Calibration ¶

from sklearn.calibration import CalibratedClassifierCV

lgbc_bal = LGBMClassifier(random_state=0)
calibrated_clf = CalibratedClassifierCV(base_estimator=lgbc_bal, cv=3, method='sigmoid')
calibrated_clf.fit(X_train_imputed, y_train)

CalibratedClassifierCV(base_estimator=LGBMClassifier(random_state=0), cv=3)

# Prediction
y_pred_calib = calibrated_clf.predict(X_test)
y_prob_pred_calib = calibrated_clf.predict_proba(X_test)[:, 1]

len(calibrated_clf.calibrated_classifiers_)

3

print("Y predicted : ", y_pred_calib)
print("Y probability predicted : ", y_prob_pred_calib[:5])

Y predicted :  [False False False ... False False  True]
Y probability predicted :  [0.01690884 0.01522014 0.017496   0.01598323 0.01401708]

# Compute Evaluation Metric
compute_evaluation_metric(calibrated_clf, X_test_imputed, y_test, y_pred_calib, y_prob_pred_calib)

 Accuracy Score :  0.976259017170725

 AUC Score :  0.9181559291804907

 Confusion Matrix : 
 [[170242    799]
 [  3407   2714]]

 Classification Report : 
               precision    recall  f1-score   support

       False       0.98      1.00      0.99    171041
        True       0.77      0.44      0.56      6121

    accuracy                           0.98    177162
   macro avg       0.88      0.72      0.78    177162
weighted avg       0.97      0.98      0.97    177162


 Concordance Index :  0.9181559244046767

 ROC curve :

 PR curve :

# Gains Table and Capture rates plot
captures(y_test, y_pred_calib, y_prob_pred_calib)

draw_calibration_curve(y_test, y_prob_pred_calib, n_bins=10)

20. Model Tuning ¶

Hyperparameter is a parameter that governs how the algorithm trains to learn the relationships. The values are set before the learning process begins.

Hyperparameter tuning refers to the automatic optimization of the hyper-parameters of a ML model.

%%time
# Define the estimator
lgbmclassifier = LGBMClassifier(random_state=0)

# Define the parameters gird
param_grid = {
    'n_estimator'   : [100,200],            # default: 100
    'num_leaves'    : [256,128],            # default: 256
    'max_depth'     : [5, 8],               # default: 8 
    'learning_rate' : [0.05, 0.1],          # default: .1
    'reg_alpha'     : [0 .1, 0.5],          # default: .5
    'class_weight'  : ['balanced', None],
}

# run grid search
grid = GridSearchCV(lgbmclassifier, param_grid=param_grid, refit = True, verbose = 3, n_jobs=-1,cv = 3)
  
# fit the model for grid search 
grid.fit(X_train, y_train)

Fitting 3 folds for each of 64 candidates, totalling 192 fits
Wall time: 2h 27min 26s

GridSearchCV(cv=3, estimator=LGBMClassifier(random_state=0), n_jobs=-1,
             param_grid={'class_weight': ['balanced', None],
                         'learning_rate': [0.05, 0.1], 'max_depth': [5, 8],
                         'n_estimator': [100, 200], 'num_leaves': [256, 128],
                         'reg_alpha': [0.1, 0.5]},
             verbose=3)

Get the best parameters corresponding to which you have best model

# Best parameter after hyper parameter tuning 
print(grid.best_params_) 
  
# Moel Parameters 
print(grid.best_estimator_)

lgbmclassifier = grid.best_estimator_

{'class_weight': None, 'learning_rate': 0.1, 'max_depth': 8, 'n_estimator': 100, 'num_leaves': 256, 'reg_alpha': 0.5}
LGBMClassifier(max_depth=8, n_estimator=100, num_leaves=256, random_state=0,
               reg_alpha=0.5)

Let's use the best model to get predictions on test dataset. We would be looking at the predicted class and predicted probability both in order to evaluate the performance of the model

# Prediction using best parameters
y_grid_pred = lgbmclassifier.predict(X_test)
y_prob_grid_pred = lgbmclassifier.predict_proba(X_test)[:, 1]
print("Y predicted : ",y_grid_pred)
print("Y probability predicted : ",y_prob_grid_pred[:5])

Y predicted :  [False False False ... False False False]
Y probability predicted :  [0.00210893 0.02470222 0.01490586 0.00864042 0.00098105]

Evaluation Metrics¶

Let's compute various evaluation metrices now

Accuracy Score
Confusion Matrix
Classification Report
AUC Score
Concodense Index
ROC curve
PR curve

# Compute Evaluation Metric
compute_evaluation_metric(lgbmclassifier, X_test, y_test, y_grid_pred, y_prob_grid_pred)

 Accuracy Score :  0.9797755726397309

 AUC Score :  0.9409700338679997

 Confusion Matrix : 
 [[170734    307]
 [  3276   2845]]

 Classification Report : 
               precision    recall  f1-score   support

       False       0.98      1.00      0.99    171041
        True       0.90      0.46      0.61      6121

    accuracy                           0.98    177162
   macro avg       0.94      0.73      0.80    177162
weighted avg       0.98      0.98      0.98    177162


 Concordance Index :  0.9409700247939532

 ROC curve :

 PR curve :

Calibration Curve¶

draw_calibration_curve(y_test, y_prob_grid_pred, n_bins=10)

Calibrate the model¶

# Calibrate
calibrated_clf = CalibratedClassifierCV(base_estimator=lgbmclassifier, cv=3)
calibrated_clf.fit(X_train, y_train)
y_pred_calib = calibrated_clf.predict(X_test)
y_prob_pred_calib = calibrated_clf.predict_proba(X_test)[:, 1]

draw_calibration_curve(y_test, y_prob_pred_calib, n_bins=10)

# Compute Evaluation Metric
compute_evaluation_metric(calibrated_clf, X_test, y_test, y_pred_calib, y_prob_pred_calib)

 Accuracy Score :  0.9801763357830686

 AUC Score :  0.9425583129340254

 Confusion Matrix : 
 [[170598    443]
 [  3069   3052]]

 Classification Report : 
               precision    recall  f1-score   support

       False       0.98      1.00      0.99    171041
        True       0.87      0.50      0.63      6121

    accuracy                           0.98    177162
   macro avg       0.93      0.75      0.81    177162
weighted avg       0.98      0.98      0.98    177162


 Concordance Index :  0.9425583124564438

 ROC curve :

 PR curve :

# Gains Table and Capture rates plot
captures(y_test, y_pred_calib, y_prob_pred_calib)

Inferences:

Accuracy score is 0.91
AUC score and Concordance index are 0.97, which are the best so far
Classfication report is also balanced between both the classes
ROC curve and PR are the best so far

Hence we can freeze the model.

21. Feature Importance ¶

Feature importance refers to techniques that assign a score to input features based on how useful they are at predicting a target variable.

lgbmclassifier = grid.best_estimator_

lgbmclassifier.feature_importances_

array([189,  60, 657, 568, 106,  58, 251,  89, 488,  10, 261, 250, 119,
       222, 197,  17,  31,  63, 146,  28,  65, 103,  62, 171,  62, 298,
       130, 193, 267, 144, 232,  96,  43, 266,  95, 195, 123,  39,  60,
        94, 286,   0,  20,  29,  66,  61,  58,   8,   8,  16,  42,  81,
         2,   0,  38,  31,   8,   1,  10,   2,  35,  14,   2,   1,  11,
        86,  64,   0,   0,   8,   9,   4,  33,   0,   0,   2,   1,   2,
         0,  33,   0,   0,   0,   0,   0,   0,   0,   0,  22,  15,   1,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   3,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   2,   0,   0,   0,   0,   1,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   1,   0,   1,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   1,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   1,   0,   0,   0,   0,   0,   0,   0,   0,   0,
         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,
        94, 172, 343,  52,   8, 232, 189, 344, 167, 188, 161, 135, 118,
       114, 177, 100, 242, 188, 178, 169, 146, 185, 186, 199, 161, 167,
       167, 136, 155, 141, 194, 184, 125, 110, 149, 186, 194, 192, 200])

feature_importance_df = pd.DataFrame({'feature' : X_train.columns, 'importance' : lgbmclassifier.feature_importances_ })

feature_importance_df = feature_importance_df.sort_values(by="importance", ascending=False)
feature_importance_df = feature_importance_df.iloc[:30,:]
feature_importance_df

plt.figure(figsize=(16, 12));
sns.barplot(x="importance", y="feature", data=feature_importance_df.sort_values(by="importance", ascending=False));
plt.title('LGB Features');

Inferences:

card1 is contributing the most in predicting if a transaction is fraud or not
card2, addr1, C13, P_emaildomain, C1 etc are some of the most important features in predicting the fraud
Certain card types, addresses and emails are at high risk of fraud, so there is a need to monitor these carefully

22. Partial Dependence and Individual Conditional Expectations (ICE) ¶

## pdp plots
from sklearn.inspection import partial_dependence, plot_partial_dependence
from sklearn.utils import validation

Fit the model

lgbmclassifier.fit(X_train, y_train)
lgbmclassifier.dummy_ = "dummy"

validation.check_is_fitted(estimator=lgbmclassifier)

Plot Partial Dependence

fig = plt.figure(figsize=(16, 12))
plot_partial_dependence(lgbmclassifier, X, ['card2'])
plt.show()

<Figure size 1152x864 with 0 Axes>

Individual Conditional Expectation (ICE) Plot - card2

plot_partial_dependence(lgbmclassifier, X, ['card2'], kind='both')

<sklearn.inspection._plot.partial_dependence.PartialDependenceDisplay at 0x11225b5bd68>

Partial Dependence and ICE Plot - C13

fig = plt.figure(figsize=(16, 12))
plot_partial_dependence(lgbmclassifier, X, ['C13'], kind='both')
plt.show()

<Figure size 1152x864 with 0 Axes>

fig = plt.figure(figsize=(16, 12))
plot_partial_dependence(lgbmclassifier, X, ['C13'])
plt.show()

<Figure size 1152x864 with 0 Axes>

23. SHAP Values ¶

SHAP values is used to reverse engineer the output of the prediction model and quantify the contribution of each predictor for a given prediction.

import shap
shap_model = shap.TreeExplainer(lgbmclassifier)
shap_values = shap_model.shap_values(X_train)

You can make a partial dependence plot using shap.dependence_plot. This shows the relationship between the feature and the Y. This also automatically includes another feature that your feature interacts frequently with.

# card2
shap.dependence_plot("card2", shap_values[0], X_train)

# card3
shap.dependence_plot("card3", shap_values[0], X_train)

Explain a single observation.

shap.initjs()  # needed to show viz
shap.force_plot(shap_model.expected_value[1], shap_values[1][14], X_train.iloc[14, :])

Add link = "logit"

shap.initjs()  # needed to show viz
shap.force_plot(shap_model.expected_value[1], shap_values[1][14], X_train.iloc[14, :], link='logit')

y_pred_calib_tr[14]

array([0.89570106, 0.10429894])

# compute SHAP values
explainer = shap.Explainer(lgbmclassifier, X_train) # , link=shap.links.logit)
shap_values_waterfall = explainer(X_train[:100])

# visualize the first prediction's explanation
shap.plots.waterfall(shap_values_waterfall[0])

Conclusion¶

The model has been trained and tested, so now one can use it to predict if any transaction would be fraud or not.

isFraud	0	1
ProductCD
C	60511	8008
H	31450	1574
R	36273	1426
S	10942	686
W	430701	8969

isFraud	0	1	odds
ProductCD
C	60511	8008	0.132340
H	31450	1574	0.050048
R	36273	1426	0.039313
S	10942	686	0.062694
W	430701	8969	0.020824

isFraud	0	1	odds	odds_ratio
ProductCD
C	60511	8008	0.132340	3.649871
H	31450	1574	0.050048	1.380295
R	36273	1426	0.039313	1.084236
S	10942	686	0.062694	1.729080
W	430701	8969	0.020824	0.574323

	prob_bin	not_fraud	fraud	perc_fraud	perc_not_fraud	cum_perc_fraud	cum_perc_not_fraud
0	Bin1((0.0472, 1.0])	12835	4882	0.797582	0.075040	79.758209	7.504049
1	Bin2((0.0215, 0.0472])	17204	512	0.083646	0.100584	88.122856	17.562456
2	Bin3((0.0132, 0.0215])	17460	256	0.041823	0.102081	92.305179	27.770535
3	Bin4((0.00914, 0.0132])	17572	144	0.023526	0.102736	94.657736	38.044095
4	Bin5((0.00659, 0.00914])	17606	109	0.017808	0.102934	96.438490	48.337533
5	Bin6((0.00477, 0.00659])	17649	68	0.011109	0.103186	97.549420	58.656112
6	Bin7((0.00333, 0.00477])	17662	54	0.008822	0.103262	98.431629	68.982291
7	Bin8((0.00219, 0.00333])	17668	48	0.007842	0.103297	99.215814	79.311978
8	Bin9((0.00121, 0.00219])	17689	27	0.004411	0.103420	99.656919	89.653943
9	Bin10((-0.0009859, 0.00121])	17696	21	0.003431	0.103461	100.000000	100.000000

	prob_bin	not_fraud	fraud	perc_fraud	perc_not_fraud	cum_perc_fraud	cum_perc_not_fraud
0	Bin1((0.0188, 0.998])	12835	4882	0.797582	0.075040	79.758209	7.504049
1	Bin2((0.0144, 0.0188])	17204	512	0.083646	0.100584	88.122856	17.562456
2	Bin3((0.0132, 0.0144])	17460	256	0.041823	0.102081	92.305179	27.770535
3	Bin4((0.0127, 0.0132])	17572	144	0.023526	0.102736	94.657736	38.044095
4	Bin5((0.0124, 0.0127])	17606	109	0.017808	0.102934	96.438490	48.337533
5	Bin6((0.0121, 0.0124])	17649	68	0.011109	0.103186	97.549420	58.656112
6	Bin7((0.0119, 0.0121])	17662	54	0.008822	0.103262	98.431629	68.982291
7	Bin8((0.0118, 0.0119])	17668	48	0.007842	0.103297	99.215814	79.311978
8	Bin9((0.0117, 0.0118])	17689	27	0.004411	0.103420	99.656919	89.653943
9	Bin10((0.010499999999999999, 0.0117])	17696	21	0.003431	0.103461	100.000000	100.000000

	TransactionID	id_01	id_02	id_03	id_04	id_05	id_06	id_07	id_08	id_09	id_10	id_11	id_12	id_13	id_14	id_15	id_16	id_17	id_18	id_19	id_20	id_21	id_22	id_23	id_24	id_25	id_26	id_27	id_28	id_29	id_30	id_31	id_32	id_33	id_34	id_35	id_36	id_37	id_38	DeviceType	DeviceInfo
601	2990046	-5.0	821814.0	NaN	NaN	0.0	0.0	NaN	NaN	NaN	NaN	100.0	NotFound	49.0	-300.0	New	NotFound	102.0	15.0	410.0	360.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	New	NotFound	iOS 11.1.2	mobile safari 11.0	32.0	2208x1242	match_status:1	T	F	F	F	mobile	iOS Device
29589	3065170	-10.0	103219.0	0.0	0.0	2.0	-5.0	NaN	NaN	0.0	0.0	100.0	Found	52.0	NaN	Found	Found	166.0	13.0	216.0	214.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	Found	Found	NaN	ie 11.0 for desktop	NaN	NaN	NaN	F	F	T	T	desktop	rv:11.0
97539	3342636	-20.0	175191.0	NaN	NaN	0.0	0.0	NaN	NaN	NaN	NaN	100.0	NotFound	33.0	NaN	New	NotFound	225.0	NaN	266.0	305.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	New	NotFound	NaN	samsung browser generic	NaN	NaN	NaN	F	F	T	F	mobile	SAMSUNG SM-G610M Build/NRD90M
141317	3561291	-5.0	135838.0	0.0	0.0	0.0	0.0	NaN	NaN	0.0	0.0	100.0	NotFound	52.0	-480.0	Found	Found	166.0	NaN	193.0	222.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	Found	Found	Windows 10	chrome 66.0	24.0	1920x1080	match_status:2	T	F	T	F	desktop	Windows
65961	3161995	-5.0	22457.0	0.0	0.0	0.0	0.0	NaN	NaN	0.0	0.0	100.0	NotFound	49.0	-360.0	Found	Found	166.0	NaN	193.0	333.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	Found	Found	Mac OS X 10_13_2	chrome 63.0	24.0	2560x1600	match_status:2	T	F	T	F	desktop	MacOS
138982	3549519	-10.0	417370.0	NaN	NaN	3.0	-23.0	NaN	NaN	NaN	NaN	100.0	NotFound	27.0	NaN	New	NotFound	225.0	15.0	290.0	127.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	New	NotFound	NaN	mobile safari 11.0	NaN	NaN	NaN	F	F	T	F	mobile	NaN

	TransactionID	TransactionDT	TransactionAmt	ProductCD	card1	card2	card3	card4	card5	card6	addr1	addr2	dist1	dist2	P_emaildomain	R_emaildomain	C1	C2	C6	C8	C9	C10	C11	C13	C14	D1	D2	D3	D4	D5	D6	D7	D8	D9	D10	D11	D12	D13	D14	D15	M1	M2	M3	M4	M5	M6	M7	M8	M9	V1	V2	V3	V4	V5	V6	V7	V8	V9	V10	V11	V12	V13	V14	V15	V16	V17	V18	V19	V20	V21	V22	V23	V24	V25	V26	V27	V28	V29	V30	V31	V32	V33	V34	V35	V36	V37	V38	V39	V40	V41	V42	V43	V44	V45	V46	V47	V48	V49	V50	V51	V52	V53	V54	V55	V56	V57	V58	V59	V60	V61	V62	V63	V64	V65	V66	V67	V68	V69	V70	V71	V72	V73	V74	V75	V76	V77	V78	V79	V80	V81	V82	V83	V84	V85	V86	V87	V88	V89	V90	V91	V92	V93	V94	V95	V96	V97	V99	V100	V101	V102	V103	V107	V108	V109	V110	V111	V112	V113	V114	V115	V116	V117	V118	V119	V120	V121	V122	V123	V124	V125	V126	V127	V128	V130	V131	V132	V133	V134	V138	V139	V140	V141	V142	V143	V144	V145	V146	V147	V148	V149	V150	V151	V152	V153	V154	V155	V156	V157	V158	V159	V160	V161	V162	V163	V164	V165	V166	V167	V168	V169	V170	V171	V172	V173	V174	V175	V176	V177	V178	V179	V180	V181	V182	V183	V184	V185	V186	V187	V188	V189	V190	V191	V192	V193	V194	V195	V196	V197	V198	V199	V200	V201	V202	V203	V204	V205	V206	V207	V208	V209	V210	V211	V212	V213	V214	V215	V216	V217	V218	V219	V220	V221	V222	V223	V224	V225	V226	V227	V228	V229	V230	V231	V232	V233	V234	V235	V236	V237	V238	V239	V240	V241	V242	V243	V244	V245	V246	V247	V248	V249	V250	V251	V252	V253	V254	V255	V256	V257	V258	V259	V260	V261	V262	V263	V264	V265	V266	V267	V268	V269	V270	V271	V272	V273	V274	V275	V276	V277	V278	V279	V280	V282	V283	V285	V287	V290	V291	V292	V293	V294	V295	V302	V303	V304	V305	V306	V307	V308	V310	V312	V316	V317	V318	V322	V323	V324	V325	V326	V327	V328	V329	V330	V331	V332	V333	V334	V335	V336	V337	V338	V339
0	2987000	86400	68.5	W	13926	NaN	150.0	discover	142.0	credit	315.0	87.0	19.0	NaN	NaN	NaN	1.0	1.0	1.0	0.0	1.0	0.0	2.0	1.0	1.0	14.0	NaN	13.0	NaN	NaN	NaN	NaN	NaN	NaN	13.0	13.0	NaN	NaN	NaN	0.0	T	T	T	M2	F	T	NaN	NaN	NaN	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	1.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	117.0	0.0	0.0	0.0	0.0	117.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	1.0	0.0	0.0	0.0	0.0	1.0	0.0	117.0	0.0	0.0	0.0	0.0	117.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
1	2987001	86401	29.0	W	2755	404.0	150.0	mastercard	102.0	credit	325.0	87.0	NaN	NaN	gmail.com	NaN	1.0	1.0	1.0	0.0	0.0	0.0	1.0	1.0	1.0	0.0	NaN	NaN	0.0	NaN	NaN	NaN	NaN	NaN	0.0	NaN	NaN	NaN	NaN	0.0	NaN	NaN	NaN	M0	T	T	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0.0	0.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2	2987002	86469	59.0	W	4663	490.0	150.0	visa	166.0	debit	330.0	87.0	287.0	NaN	outlook.com	NaN	1.0	1.0	1.0	0.0	1.0	0.0	1.0	1.0	1.0	0.0	NaN	NaN	0.0	NaN	NaN	NaN	NaN	NaN	0.0	315.0	NaN	NaN	NaN	315.0	T	T	T	M0	F	F	F	F	F	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
3	2987003	86499	50.0	W	18132	567.0	150.0	mastercard	117.0	debit	476.0	87.0	NaN	NaN	yahoo.com	NaN	2.0	5.0	4.0	0.0	1.0	0.0	1.0	25.0	1.0	112.0	112.0	0.0	94.0	0.0	NaN	NaN	NaN	NaN	84.0	NaN	NaN	NaN	NaN	111.0	NaN	NaN	NaN	M0	T	F	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	48.0	28.0	10.0	4.0	1.0	38.0	24.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	50.0	1758.0	925.0	354.0	135.0	50.0	1404.0	790.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1.0	28.0	0.0	0.0	10.0	4.0	1.0	1.0	1.0	1.0	38.0	24.0	0.0	0.0	0.0	1.0	50.0	1758.0	925.0	354.0	135.0	50.0	1404.0	790.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
4	2987004	86506	50.0	H	4497	514.0	150.0	mastercard	102.0	credit	420.0	87.0	NaN	NaN	gmail.com	NaN	1.0	1.0	1.0	1.0	0.0	1.0	1.0	1.0	1.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	6.0	18.0	140.0	0.0	0.0	0.0	0.0	1803.0	49.0	64.0	0.0	0.0	0.0	0.0	0.0	0.0	15557.990234	169690.796875	0.0	0.0	0.0	515.0	5155.0	2840.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0

	TransactionID	id_01	id_02	id_03	id_04	id_05	id_06	id_07	id_08	id_09	id_10	id_11	id_12	id_13	id_14	id_15	id_16	id_17	id_18	id_19	id_20	id_21	id_22	id_23	id_24	id_25	id_26	id_27	id_28	id_29	id_30	id_31	id_32	id_33	id_34	id_35	id_36	id_37	id_38	DeviceType	DeviceInfo
0	2987004	0.0	70787.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	100.0	NotFound	NaN	-480.0	New	NotFound	166.0	NaN	542.0	144.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	New	NotFound	Android 7.0	samsung browser 6.2	32.0	2220x1080	match_status:2	T	F	T	T	mobile	SAMSUNG SM-G892A Build/NRD90M
1	2987008	-5.0	98945.0	NaN	NaN	0.0	-5.0	NaN	NaN	NaN	NaN	100.0	NotFound	49.0	-300.0	New	NotFound	166.0	NaN	621.0	500.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	New	NotFound	iOS 11.1.2	mobile safari 11.0	32.0	1334x750	match_status:1	T	F	F	T	mobile	iOS Device
2	2987010	-5.0	191631.0	0.0	0.0	0.0	0.0	NaN	NaN	0.0	0.0	100.0	NotFound	52.0	NaN	Found	Found	121.0	NaN	410.0	142.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	Found	Found	NaN	chrome 62.0	NaN	NaN	NaN	F	F	T	T	desktop	Windows
3	2987011	-5.0	221832.0	NaN	NaN	0.0	-6.0	NaN	NaN	NaN	NaN	100.0	NotFound	52.0	NaN	New	NotFound	225.0	NaN	176.0	507.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	New	NotFound	NaN	chrome 62.0	NaN	NaN	NaN	F	F	T	T	desktop	NaN
4	2987016	0.0	7460.0	0.0	0.0	1.0	0.0	NaN	NaN	0.0	0.0	100.0	NotFound	NaN	-300.0	Found	Found	166.0	15.0	529.0	575.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	Found	Found	Mac OS X 10_11_6	chrome 62.0	24.0	1280x800	match_status:2	T	F	T	T	desktop	MacOS

	TransactionID	id_01	id_02	id_03	id_04	id_05	id_06	id_07	id_08	id_09	id_10	id_11	id_12	id_13	id_14	id_15	id_16	id_17	id_18	id_19	id_20	id_21	id_22	id_23	id_24	id_25	id_26	id_27	id_28	id_29	id_30	id_31	id_32	id_33	id_34	id_35	id_36	id_37	id_38	DeviceType	DeviceInfo
count	144233	144233	140872	66324	66324	136865	136865	5155	5155	74926	74926	140978	NaN	127320	80044	NaN	NaN	139369	45113	139318	139261	5159	5169	NaN	4747	5132	5163	NaN	NaN	NaN	NaN	NaN	77586	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
mean	3.23633e+06	NaN	174717	0	-0	NaN	NaN	inf	-inf	0	-0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	inf	NaN	NaN	inf	inf	NaN	12.7891	inf	inf	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
std	178850	0	159652	0	0	0	0	11.3828	26.0781	0	0	0	NaN	0	NaN	NaN	NaN	0	1.56152	NaN	NaN	inf	6.89844	NaN	2.37109	97.4375	32.0938	NaN	NaN	NaN	NaN	NaN	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
min	2.987e+06	-100	1	-13	-28	-72	-100	-46	-100	-36	-100	90	NaN	10	-660	NaN	NaN	100	10	100	100	100	10	NaN	11	100	100	NaN	NaN	NaN	NaN	NaN	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
25%	3.07714e+06	-10	67992	0	0	0	-6	5	-48	0	0	100	NaN	49	-360	NaN	NaN	166	13	266	256	252	14	NaN	11	321	119	NaN	NaN	NaN	NaN	NaN	24	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
50%	3.19882e+06	-5	125800	0	0	0	0	14	-34	0	0	100	NaN	52	-300	NaN	NaN	166	15	341	472	252	14	NaN	11	321	149	NaN	NaN	NaN	NaN	NaN	24	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
75%	3.39292e+06	-5	228749	0	0	1	0	22	-23	0	0	100	NaN	52	-300	NaN	NaN	225	15	427	533	486.5	14	NaN	15	371	169	NaN	NaN	NaN	NaN	NaN	32	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
max	3.57753e+06	0	999595	10	0	52	0	61	0	25	0	100	NaN	64	720	NaN	NaN	229	29	671	661	854	44	NaN	26	548	216	NaN	NaN	NaN	NaN	NaN	32	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
counts	144233	144233	140872	66324	66324	136865	136865	5155	5155	74926	74926	140978	144233	127320	80044	140985	129340	139369	45113	139318	139261	5159	5169	5169	4747	5132	5163	5169	140978	140978	77565	140282	77586	73289	77805	140985	140985	140985	140985	140810	118666
uniques	144233	77	115655	24	15	93	101	84	94	46	62	146	2	54	25	3	2	104	18	522	394	490	25	3	12	341	95	2	2	2	75	130	4	260	4	2	2	2	2	2	1786
missing	0	0	3361	77909	77909	7368	7368	139078	139078	69307	69307	3255	0	16913	64189	3248	14893	4864	99120	4915	4972	139074	139064	139064	139486	139101	139070	139064	3255	3255	66668	3951	66647	70944	66428	3248	3248	3248	3248	3423	25567
missing_perc	0%	0%	2.33%	54.02%	54.02%	5.11%	5.11%	96.43%	96.43%	48.05%	48.05%	2.26%	0%	11.73%	44.50%	2.25%	10.33%	3.37%	68.72%	3.41%	3.45%	96.42%	96.42%	96.42%	96.71%	96.44%	96.42%	96.42%	2.26%	2.26%	46.22%	2.74%	46.21%	49.19%	46.06%	2.25%	2.25%	2.25%	2.25%	2.37%	17.73%
types	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	bool	numeric	numeric	categorical	bool	numeric	numeric	numeric	numeric	numeric	numeric	categorical	numeric	numeric	numeric	bool	bool	bool	categorical	categorical	numeric	categorical	categorical	bool	bool	bool	bool	bool	categorical

	TransactionID	isFraud	TransactionDT	TransactionAmt	ProductCD	card1	card2	card3	card4	card5	card6	addr1	addr2	dist1	dist2	P_emaildomain	R_emaildomain	C1	C2	C3	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	D1	D2	D3	D4	D5	D6	D7	D8	D9	D10	D11	D12	D13	D14	D15	M1	M2	M3	M4	M5	M6	M7	M8	M9	V1	V2	V3	V4	V5	V6	V7	V8	V9	V10	V11	V12	V13	V14	V15	V16	V17	V18	V19	V20	V21	V22	V23	V24	V25	V26	V27	V28	V29	V30	V31	V32	V33	V34	V35	V36	V37	V38	V39	V40	V41	V42	V43	V44	V45	V46	V47	V48	V49	V50	V51	V52	V53	V54	V55	V56	V57	V58	V59	V60	V61	V62	V63	V64	V65	V66	V67	V68	V69	V70	V71	V72	V73	V74	V75	V76	V77	V78	V79	V80	V81	V82	V83	V84	V85	V86	V87	V88	V89	V90	V91	V92	V93	V94	V95	V96	V97	V98	V99	V100	V101	V102	V103	V104	V105	V106	V107	V108	V109	V110	V111	V112	V113	V114	V115	V116	V117	V118	V119	V120	V121	V122	V123	V124	V125	V126	V127	V128	V129	V130	V131	V132	V133	V134	V135	V136	V137	V138	V139	V140	V141	V142	V143	V144	V145	V146	V147	V148	V149	V150	V151	V152	V153	V154	V155	V156	V157	V158	V159	V160	V161	V162	V163	V164	V165	V166	V167	V168	V169	V170	V171	V172	V173	V174	V175	V176	V177	V178	V179	V180	V181	V182	V183	V184	V185	V186	V187	V188	V189	V190	V191	V192	V193	V194	V195	V196	V197	V198	V199	V200	V201	V202	V203	V204	V205	V206	V207	V208	V209	V210	V211	V212	V213	V214	V215	V216	V217	V218	V219	V220	V221	V222	V223	V224	V225	V226	V227	V228	V229	V230	V231	V232	V233	V234	V235	V236	V237	V238	V239	V240	V241	V242	V243	V244	V245	V246	V247	V248	V249	V250	V251	V252	V253	V254	V255	V256	V257	V258	V259	V260	V261	V262	V263	V264	V265	V266	V267	V268	V269	V270	V271	V272	V273	V274	V275	V276	V277	V278	V279	V280	V281	V282	V283	V284	V285	V286	V287	V288	V289	V290	V291	V292	V293	V294	V295	V296	V297	V298	V299	V300	V301	V302	V303	V304	V305	V306	V307	V308	V309	V310	V311	V312	V313	V314	V315	V316	V317	V318	V319	V320	V321	V322	V323	V324	V325	V326	V327	V328	V329	V330	V331	V332	V333	V334	V335	V336	V337	V338	V339
count	590540	590540	590540	590540	NaN	590540	581607	588975	NaN	586281	NaN	524834	524834	238269	37627	NaN	NaN	590540	590540	590540	590540	590540	590540	590540	590540	590540	590540	590540	590540	590540	590540	589271	309743	327662	421618	280699	73187	38917	74926	74926	514518	311253	64717	61952	62187	501427	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	311253	311253	311253	311253	311253	311253	311253	311253	311253	311253	311253	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	81945	81945	81945	81945	81945	81951	81951	81951	81945	81945	81945	81945	81951	81951	81951	81945	81945	81945	81945	81945	81945	81951	81951	81945	81945	81945	81951	81951	81951	139631	139631	139819	139819	139819	139631	139631	139819	139819	139631	139631	139631	139631	139819	139631	139631	139631	139819	139819	139631	139631	139819	139819	139631	139631	139631	139631	139819	139819	139631	139819	139819	139631	139819	139819	139631	139631	139631	139631	139631	139631	139819	139819	139819	139631	139631	139631	139631	139631	139631	130430	130430	130430	141416	141416	141416	130430	130430	130430	130430	141416	130430	130430	130430	130430	130430	130430	141416	130430	130430	130430	141416	141416	130430	130430	130430	130430	130430	141416	130430	130430	130430	130430	141416	141416	130430	130430	130430	141416	141416	130430	130430	141416	130430	130430	130430	130430	130430	130430	130430	130430	130430	130430	141416	141416	141416	130430	130430	130430	130430	130430	130430	590528	590528	589271	589271	589271	590528	590528	590528	590528	589271	589271	590528	590528	590528	590528	590528	590528	589271	590528	590528	590528	589271	589271	590528	590528	590528	590528	590528	590528	590528	590528	590528	590528	590528	589271	589271	589271	590528	590528	590528	590528	590528	590528	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351
mean	3.28227e+06	0.03499	7.37231e+06	NaN	NaN	9898.73	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	inf	NaN	NaN	NaN	NaN	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	inf	NaN	0	NaN	NaN	inf	inf	inf	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0	NaN	NaN	NaN	NaN	NaN	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	129.979	336.612	204.094	NaN	NaN	NaN	103.513	204.889	145.972	17.2501	38.8212	26.3651	0	NaN	NaN	0	0	NaN	NaN	NaN	0	0	0	0	NaN	NaN	NaN	0	0	0	0	NaN	NaN	NaN	47453.2	NaN	NaN	NaN	877.889	2239.91	359.469	NaN	NaN	0	NaN	NaN	0	0	0	0	NaN	NaN	NaN	NaN	NaN	0	NaN	NaN	0	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	444.147	1078.33	686.957	NaN	NaN	NaN	NaN	NaN	NaN	385.137	765.988	536.303	38.4375	133.208	71.1071	NaN	NaN	NaN	0	NaN	NaN	0	0	0	0	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0	0	0	0	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	117.391	201.658	153.521	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	107.152	NaN	31.7973	51.9566	42.3282	NaN	NaN	0	NaN	NaN	0	NaN	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0	NaN	NaN	0	0	NaN	NaN	NaN	NaN	139.749	408.682	230.413	NaN	NaN	NaN	NaN	NaN	NaN	NaN	109.819	247.607	162.153	18.3725	42.0731	28.3266	NaN	NaN	NaN	0	NaN	0	0	NaN	0	721.742	1375.78	1014.62	NaN	NaN	NaN	55.3524	151.161	100.701
std	170474	0.183755	4.61722e+06	NaN	NaN	4901.17	NaN	0	NaN	0	NaN	NaN	0	NaN	inf	NaN	NaN	NaN	NaN	0	NaN	NaN	NaN	NaN	NaN	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	inf	NaN	0	NaN	NaN	inf	inf	inf	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	NaN	NaN	NaN	0	0	0	NaN	NaN	NaN	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	2346.95	4238.67	3010.26	NaN	NaN	NaN	2266.11	3796.32	2772.99	293.848	451.808	348.333	0	0	0	0	0	NaN	0	NaN	0	0	0	0	NaN	0	0	0	0	0	0	0	0	NaN	142076	NaN	NaN	NaN	6049.17	8223.26	1244.46	NaN	NaN	0	0	0	0	0	0	0	0	NaN	NaN	NaN	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	4683.83	9105.61	6048.98	NaN	NaN	NaN	NaN	NaN	NaN	4541.84	7496.12	5471.66	571.834	1040.45	680.268	NaN	NaN	NaN	0	NaN	NaN	0	0	0	0	NaN	0	0	0	NaN	NaN	NaN	0	0	0	0	0	0	0	0	0	0	0	NaN	0	0	0	0	0	0	0	0	0	0	0	0	0	NaN	0	0	0	1294.85	2284.83	1605.51	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1258.73	NaN	615.66	732.145	660.612	NaN	NaN	0	0	0	0	0	0	0	0	0	0	NaN	NaN	NaN	NaN	NaN	0	0	0	0	0	0	0	0	0	0	2348.85	4391.99	3021.92	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2270.03	3980.04	2793.34	332.305	473.499	382.053	NaN	NaN	NaN	0	0	0	0	0	0	6217.22	11169.3	7955.74	NaN	NaN	NaN	668.487	1095.03	814.947
min	2.987e+06	0	86400	0.250977	NaN	1000	100	100	NaN	100	NaN	100	10	0	0	NaN	NaN	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	-122	0	-83	0	0	0	0	-53	-83	0	-193	-83	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
25%	3.13463e+06	0	3.02706e+06	43.3125	NaN	6019	214	150	NaN	166	NaN	204	87	3	7	NaN	NaN	1	1	0	0	0	1	0	0	0	0	1	0	1	1	0	26	1	0	1	0	0	0.958496	0.208374	0	0	0	0	0	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1	1	1	1	1	1	1	1	1	0	0	0	0	1	0	0	0	0	1	1	0	0	1	1	1	1	0	0	0	0	0	0	0	0	0	0	1	1	0	0	1	0	0	1	1	1	1	0	0	0	0	0	0	0	1	1	0	0	0	0	1	1	0	0	1	1	1	0	0	0	0	0	0	0	0	0	1	1	0	0	0	1	1	0	0	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	0	0	0	0	0	1	1	1	0	0	0	0	0	0	0	0	0	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
50%	3.28227e+06	0	7.30653e+06	68.75	NaN	9678	361	150	NaN	226	NaN	299	87	8	37	NaN	NaN	1	1	0	0	0	1	0	0	1	0	1	0	3	1	3	97	8	26	10	0	0	37.875	0.666504	15	43	0	0	0	52	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1	1	1	1	1	1	1	1	1	0	0	1	1	1	0	0	0	0	1	1	0	0	1	1	1	1	0	0	0	0	0	0	0	0	1	1	1	1	0	0	1	0	0	1	1	1	1	0	0	0	0	0	1	1	1	1	0	0	0	0	1	1	0	0	1	1	1	0	0	0	0	0	0	0	1	1	1	1	0	0	0	1	1	0	0	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	0	0	0	0	0	0	0	1	1	1	1	1	1	1	1	1	1	1	0	0	0	0	0	0	0	0	0	0	0	1	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	0	0	0	0	0	1	1	1	0	0	0	0	0	0	0	0	0	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	0	0	0	0	0	0	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
75%	3.4299e+06	0	1.12466e+07	125	NaN	14184	512	150	NaN	226	NaN	330	87	24	206	NaN	NaN	3	3	0	0	1	2	0	0	2	0	2	0	12	2	122	276	27	253	32	40	17	188	0.833496	197	274	13	0	2	314	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	0	0	0	1	1	0	0	1	1	1	1	0	0	1	1	0	0	0	0	1	1	1	1	0	0	1	0	0	1	1	1	1	1	1	0	0	0	1	1	1	1	0	0	0	0	1	1	0	0	1	1	1	0	1	1	0	0	0	0	1	1	1	1	0	0	0	1	1	0	0	1	1	1	0	1	1	0	0	0	0	1	0	0	1	0	0	0	0	0	0	0	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	107.95	0	0	59	0	0	0	0	0	0	0	0	1	1	0	0	0	0	1	0	0	1	1	1	1	1	1	1	1	1	1	1	0	0	0	0	0	0	0	0	0	1	0	1	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	30.9244	20	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	0	1	1	0	0	0	0	0	1	1	1	0	0	0	0	0	0	0	0	0	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	33.5935	20.8975	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0	1	1	0	1	0	0	0	0	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	1	0	151.381	35.97	0	107.938	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	25	0	0	0	0	0	0	0
max	3.57754e+06	1	1.58111e+07	31936	NaN	18396	600	231	NaN	237	NaN	540	102	10288	11624	NaN	NaN	4684	5692	26	2252	349	2252	2256	3332	210	3256	3188	3188	2918	1429	640	640	819	869	819	873	843	1708	0.958496	876	670	648	847	878	879	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1	8	9	6	6	9	9	8	8	4	5	3	6	1	7	15	15	15	7	15	5	8	13	13	7	13	4	4	5	9	7	15	7	13	3	5	54	54	15	24	1	8	8	48	48	6	12	5	5	5	6	12	5	6	17	51	6	10	16	16	6	10	7	7	1	7	8	2	5	6	6	10	7	8	4	6	30	31	7	19	19	7	7	7	7	30	30	1	2	5	6	7	7	2	880	1410	976	12	88	28	869	1285	928	15	99	55	1	7	7	7	9	9	9	6	6	6	3	3	3	3	3	3	13	13	13	160000	160000	160000	55136	55136	55136	93736	133915	98476	90750	90750	90750	22	33	33	5	9	869	62	297	24	26	20	20	3388	57	69	18	18	24	24	24	24	55136	641511	3300	3300	3300	93736	98476	104060	872	964	19	48	61	31	7	8	14	48	861	1235	920	83	24	83	41	16	31	38	218	30	30	42	21	44	37	7	16	38	14	21	45	45	55	104060	139777	104060	55136	55136	55136	3300	8048	3300	92888	129006	97628	104060	104060	104060	303	400	378	25	384	384	16	144	51	242	360	54	176	65	293	337	332	121	23	45	39	23	23	7	5	20	57	22	262	45	18	36	22	18	18	24	163	60	87	87	48	66	285	8	49	20	153600	153600	153600	55136	55136	55136	55136	4000	4000	4000	51200	66000	51200	104060	104060	104060	880	975	22	32	68	12	95	8	31	10	12	67	1055	323	869	1286	928	93	12	93	49	11	13	16	20	16	2	108800	145765	108800	55136	55136	55136	55136	4816	7520	4816	93736	134021	98476	104060	104060	104060	880	1411	976	12	44	18	15	99	55	160000	160000	160000	55136	55136	55136	104060	104060	104060
counts	590540	590540	590540	590540	590540	590540	581607	588975	588963	586281	588969	524834	524834	238269	37627	496084	137291	590540	590540	590540	590540	590540	590540	590540	590540	590540	590540	590540	590540	590540	590540	589271	309743	327662	421618	280699	73187	38917	74926	74926	514518	311253	64717	61952	62187	501427	319440	319440	319440	309096	240058	421180	244275	244288	244288	311253	311253	311253	311253	311253	311253	311253	311253	311253	311253	311253	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	514467	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	421571	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	513444	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	501376	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	590226	81945	81945	81945	81945	81945	81951	81951	81951	81945	81945	81945	81945	81951	81951	81951	81945	81945	81945	81945	81945	81945	81951	81951	81945	81945	81945	81951	81951	81951	139631	139631	139819	139819	139819	139631	139631	139819	139819	139631	139631	139631	139631	139819	139631	139631	139631	139819	139819	139631	139631	139819	139819	139631	139631	139631	139631	139819	139819	139631	139819	139819	139631	139819	139819	139631	139631	139631	139631	139631	139631	139819	139819	139819	139631	139631	139631	139631	139631	139631	130430	130430	130430	141416	141416	141416	130430	130430	130430	130430	141416	130430	130430	130430	130430	130430	130430	141416	130430	130430	130430	141416	141416	130430	130430	130430	130430	130430	141416	130430	130430	130430	130430	141416	141416	130430	130430	130430	141416	141416	130430	130430	141416	130430	130430	130430	130430	130430	130430	130430	130430	130430	130430	141416	141416	141416	130430	130430	130430	130430	130430	130430	590528	590528	589271	589271	589271	590528	590528	590528	590528	589271	589271	590528	590528	590528	590528	590528	590528	589271	590528	590528	590528	589271	589271	590528	590528	590528	590528	590528	590528	590528	590528	590528	590528	590528	589271	589271	589271	590528	590528	590528	590528	590528	590528	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351	82351
uniques	590540	2	573349	8195	5	13553	500	114	4	119	4	332	74	2412	1699	59	60	1495	1167	27	1223	319	1291	1069	1130	205	1122	1343	1066	1464	1108	641	641	649	808	688	829	597	5367	24	818	676	635	577	802	859	2	2	2	3	2	2	2	2	2	2	9	10	7	7	10	10	9	9	5	6	4	7	2	8	15	16	16	8	15	6	9	14	14	7	13	4	4	6	8	8	15	7	13	4	6	55	55	16	18	2	9	9	49	49	7	9	6	6	6	7	9	6	7	18	52	7	11	17	17	7	11	8	8	2	8	9	3	6	7	7	11	8	9	5	7	31	32	8	20	20	8	8	8	8	31	31	2	3	6	7	8	8	3	881	1410	976	13	89	29	870	1285	928	16	100	56	2	8	8	8	10	10	10	7	7	7	4	4	4	4	4	4	14	14	14	10299	24414	14507	1608	5511	3097	6560	9949	8178	3724	4852	4252	23	34	34	6	10	870	63	260	25	27	21	21	1344	56	39	19	19	25	25	25	25	2492	9621	79	185	106	1978	2547	987	873	965	20	49	62	32	8	9	15	49	862	1236	921	84	25	84	42	17	32	39	215	31	31	43	22	45	38	8	17	39	15	22	46	46	56	10970	14951	12858	1953	1581	2705	2093	2674	2262	7624	8868	8317	2282	2747	2532	304	401	379	26	77	76	17	79	35	81	50	55	91	66	294	338	333	122	24	46	40	24	24	6	5	21	43	23	58	46	19	23	23	19	19	25	66	45	46	48	49	67	68	9	41	21	10422	13358	11757	1871	2884	2286	151	1972	2286	2082	4689	8315	4965	2263	2540	2398	881	975	23	33	62	13	96	9	32	11	13	58	219	173	870	1286	928	94	13	94	50	12	14	17	21	17	2	16210	37367	23064	3239	7759	2526	5143	3915	5974	4540	9814	15184	12309	4799	6439	5560	881	1411	976	13	45	19	16	100	56	1758	2453	1971	143	669	355	254	380	334
missing	0	0	0	0	0	0	8933	1565	1577	4259	1571	65706	65706	352271	552913	94456	453249	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1269	280797	262878	168922	309841	517353	551623	515614	515614	76022	279287	525823	528588	528353	89113	271100	271100	271100	281444	350482	169360	346265	346252	346252	279287	279287	279287	279287	279287	279287	279287	279287	279287	279287	279287	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	76073	168969	168969	168969	168969	168969	168969	168969	168969	168969	168969	168969	168969	168969	168969	168969	168969	168969	168969	77096	77096	77096	77096	77096	77096	77096	77096	77096	77096	77096	77096	77096	77096	77096	77096	77096	77096	77096	77096	77096	77096	89164	89164	89164	89164	89164	89164	89164	89164	89164	89164	89164	89164	89164	89164	89164	89164	89164	89164	89164	89164	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	314	508595	508595	508595	508595	508595	508589	508589	508589	508595	508595	508595	508595	508589	508589	508589	508595	508595	508595	508595	508595	508595	508589	508589	508595	508595	508595	508589	508589	508589	450909	450909	450721	450721	450721	450909	450909	450721	450721	450909	450909	450909	450909	450721	450909	450909	450909	450721	450721	450909	450909	450721	450721	450909	450909	450909	450909	450721	450721	450909	450721	450721	450909	450721	450721	450909	450909	450909	450909	450909	450909	450721	450721	450721	450909	450909	450909	450909	450909	450909	460110	460110	460110	449124	449124	449124	460110	460110	460110	460110	449124	460110	460110	460110	460110	460110	460110	449124	460110	460110	460110	449124	449124	460110	460110	460110	460110	460110	449124	460110	460110	460110	460110	449124	449124	460110	460110	460110	449124	449124	460110	460110	449124	460110	460110	460110	460110	460110	460110	460110	460110	460110	460110	449124	449124	449124	460110	460110	460110	460110	460110	460110	12	12	1269	1269	1269	12	12	12	12	1269	1269	12	12	12	12	12	12	1269	12	12	12	1269	1269	12	12	12	12	12	12	12	12	12	12	12	1269	1269	1269	12	12	12	12	12	12	508189	508189	508189	508189	508189	508189	508189	508189	508189	508189	508189	508189	508189	508189	508189	508189	508189	508189
missing_perc	0%	0%	0%	0%	0%	0%	1.51%	0.27%	0.27%	0.72%	0.27%	11.13%	11.13%	59.65%	93.63%	15.99%	76.75%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0%	0.21%	47.55%	44.51%	28.60%	52.47%	87.61%	93.41%	87.31%	87.31%	12.87%	47.29%	89.04%	89.51%	89.47%	15.09%	45.91%	45.91%	45.91%	47.66%	59.35%	28.68%	58.64%	58.63%	58.63%	47.29%	47.29%	47.29%	47.29%	47.29%	47.29%	47.29%	47.29%	47.29%	47.29%	47.29%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	12.88%	28.61%	28.61%	28.61%	28.61%	28.61%	28.61%	28.61%	28.61%	28.61%	28.61%	28.61%	28.61%	28.61%	28.61%	28.61%	28.61%	28.61%	28.61%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	13.06%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	15.10%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	0.05%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	86.12%	76.36%	76.36%	76.32%	76.32%	76.32%	76.36%	76.36%	76.32%	76.32%	76.36%	76.36%	76.36%	76.36%	76.32%	76.36%	76.36%	76.36%	76.32%	76.32%	76.36%	76.36%	76.32%	76.32%	76.36%	76.36%	76.36%	76.36%	76.32%	76.32%	76.36%	76.32%	76.32%	76.36%	76.32%	76.32%	76.36%	76.36%	76.36%	76.36%	76.36%	76.36%	76.32%	76.32%	76.32%	76.36%	76.36%	76.36%	76.36%	76.36%	76.36%	77.91%	77.91%	77.91%	76.05%	76.05%	76.05%	77.91%	77.91%	77.91%	77.91%	76.05%	77.91%	77.91%	77.91%	77.91%	77.91%	77.91%	76.05%	77.91%	77.91%	77.91%	76.05%	76.05%	77.91%	77.91%	77.91%	77.91%	77.91%	76.05%	77.91%	77.91%	77.91%	77.91%	76.05%	76.05%	77.91%	77.91%	77.91%	76.05%	76.05%	77.91%	77.91%	76.05%	77.91%	77.91%	77.91%	77.91%	77.91%	77.91%	77.91%	77.91%	77.91%	77.91%	76.05%	76.05%	76.05%	77.91%	77.91%	77.91%	77.91%	77.91%	77.91%	0.00%	0.00%	0.21%	0.21%	0.21%	0.00%	0.00%	0.00%	0.00%	0.21%	0.21%	0.00%	0.00%	0.00%	0.00%	0.00%	0.00%	0.21%	0.00%	0.00%	0.00%	0.21%	0.21%	0.00%	0.00%	0.00%	0.00%	0.00%	0.00%	0.00%	0.00%	0.00%	0.00%	0.00%	0.21%	0.21%	0.21%	0.00%	0.00%	0.00%	0.00%	0.00%	0.00%	86.05%	86.05%	86.05%	86.05%	86.05%	86.05%	86.05%	86.05%	86.05%	86.05%	86.05%	86.05%	86.05%	86.05%	86.05%	86.05%	86.05%	86.05%
types	numeric	bool	numeric	numeric	categorical	numeric	numeric	numeric	categorical	numeric	categorical	numeric	numeric	numeric	numeric	categorical	categorical	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	bool	bool	bool	categorical	bool	bool	bool	bool	bool	bool	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	bool	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	bool	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	bool	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	bool	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	bool	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	bool	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric	numeric

	TransactionID	TransactionDT	TransactionAmt	ProductCD	card1	card2	card3	card4	card5	card6	addr1	addr2	dist1	P_emaildomain	R_emaildomain	C1	C2	C6	C8	C9	C10	C11	C13	C14	D1	D2	D3	D4	D5	D6	D8	D9	D10	D11	D12	D13	D14	D15	M1	M2	M3	M4	M5	M6	M7	M8	M9	V1	V2	V3	V4	V5	V6	V7	V8	V9	V10	V11	V12	V13	V14	V15	V16	V17	V18	V19	V20	V21	V22	V23	V24	V25	V26	V27	V28	V29	V30	V31	V32	V33	V34	V35	V36	V37	V38	V39	V40	V41	V42	V43	V44	V45	V46	V47	V48	V49	V50	V51	V52	V53	V54	V55	V56	V57	V58	V59	V60	V61	V62	V63	V64	V65	V66	V67	V68	V69	V70	V71	V72	V73	V74	V75	V76	V77	V78	V79	V80	V81	V82	V83	V84	V85	V86	V87	V88	V89	V90	V91	V92	V93	V94	V95	V96	V97	V99	V100	V101	V102	V103	V107	V108	V109	V110	V111	V112	V113	V114	V115	V116	V117	V118	V119	V120	V121	V122	V123	V124	V125	V126	V127	V128	V130	V131	V132	V133	V134	V138	V139	V140	V141	V142	V143	V144	V145	V146	V147	V148	V149	V150	V151	V152	V153	V154	V155	V156	V157	V158	V159	V160	V161	V162	V163	V164	V165	V166	V167	V168	V169	V170	V171	V172	V173	V174	V175	V176	V177	V178	V179	V180	V181	V182	V183	V184	V185	V186	V187	V188	V189	V190	V191	V192	V193	V194	V195	V196	V197	...	V140_missing_flag	V141_missing_flag	V142_missing_flag	V143_missing_flag	V144_missing_flag	V145_missing_flag	V146_missing_flag	V147_missing_flag	V148_missing_flag	V149_missing_flag	V150_missing_flag	V151_missing_flag	V152_missing_flag	V153_missing_flag	V154_missing_flag	V155_missing_flag	V156_missing_flag	V157_missing_flag	V158_missing_flag	V159_missing_flag	V160_missing_flag	V161_missing_flag	V162_missing_flag	V163_missing_flag	V164_missing_flag	V165_missing_flag	V166_missing_flag	V167_missing_flag	V168_missing_flag	V169_missing_flag	V170_missing_flag	V171_missing_flag	V172_missing_flag	V173_missing_flag	V174_missing_flag	V175_missing_flag	V176_missing_flag	V177_missing_flag	V178_missing_flag	V179_missing_flag	V180_missing_flag	V181_missing_flag	V182_missing_flag	V183_missing_flag	V184_missing_flag	V185_missing_flag	V186_missing_flag	V187_missing_flag	V188_missing_flag	V189_missing_flag	V190_missing_flag	V191_missing_flag	V192_missing_flag	V193_missing_flag	V194_missing_flag	V195_missing_flag	V196_missing_flag	V197_missing_flag	V198_missing_flag	V199_missing_flag	V200_missing_flag	V201_missing_flag	V202_missing_flag	V203_missing_flag	V204_missing_flag	V205_missing_flag	V206_missing_flag	V207_missing_flag	V208_missing_flag	V209_missing_flag	V210_missing_flag	V211_missing_flag	V212_missing_flag	V213_missing_flag	V214_missing_flag	V215_missing_flag	V216_missing_flag	V217_missing_flag	V218_missing_flag	V219_missing_flag	V220_missing_flag	V221_missing_flag	V222_missing_flag	V223_missing_flag	V224_missing_flag	V225_missing_flag	V226_missing_flag	V227_missing_flag	V228_missing_flag	V229_missing_flag	V230_missing_flag	V231_missing_flag	V232_missing_flag	V233_missing_flag	V234_missing_flag	V235_missing_flag	V236_missing_flag	V237_missing_flag	V238_missing_flag	V239_missing_flag	V240_missing_flag	V241_missing_flag	V242_missing_flag	V243_missing_flag	V244_missing_flag	V245_missing_flag	V246_missing_flag	V247_missing_flag	V248_missing_flag	V249_missing_flag	V250_missing_flag	V251_missing_flag	V252_missing_flag	V253_missing_flag	V254_missing_flag	V255_missing_flag	V256_missing_flag	V257_missing_flag	V258_missing_flag	V259_missing_flag	V260_missing_flag	V261_missing_flag	V262_missing_flag	V263_missing_flag	V264_missing_flag	V265_missing_flag	V266_missing_flag	V267_missing_flag	V268_missing_flag	V269_missing_flag	V270_missing_flag	V271_missing_flag	V272_missing_flag	V273_missing_flag	V274_missing_flag	V275_missing_flag	V276_missing_flag	V277_missing_flag	V278_missing_flag	V279_missing_flag	V280_missing_flag	V281_missing_flag	V282_missing_flag	V283_missing_flag	V284_missing_flag	V285_missing_flag	V286_missing_flag	V287_missing_flag	V288_missing_flag	V289_missing_flag	V290_missing_flag	V291_missing_flag	V292_missing_flag	V293_missing_flag	V294_missing_flag	V295_missing_flag	V296_missing_flag	V297_missing_flag	V298_missing_flag	V299_missing_flag	V300_missing_flag	V301_missing_flag	V302_missing_flag	V303_missing_flag	V304_missing_flag	V305_missing_flag	V306_missing_flag	V307_missing_flag	V308_missing_flag	V309_missing_flag	V310_missing_flag	V311_missing_flag	V312_missing_flag	V313_missing_flag	V314_missing_flag	V315_missing_flag	V316_missing_flag	V317_missing_flag	V318_missing_flag	V319_missing_flag	V320_missing_flag	V321_missing_flag	V322_missing_flag	V323_missing_flag	V324_missing_flag	V325_missing_flag	V326_missing_flag	V327_missing_flag	V328_missing_flag	V329_missing_flag	V330_missing_flag	V331_missing_flag	V332_missing_flag	V333_missing_flag	V334_missing_flag	V335_missing_flag	V336_missing_flag	V337_missing_flag	V338_missing_flag	V339_missing_flag	id_01_missing_flag	id_02_missing_flag	id_03_missing_flag	id_04_missing_flag	id_05_missing_flag	id_06_missing_flag	id_07_missing_flag	id_08_missing_flag	id_09_missing_flag	id_10_missing_flag	id_11_missing_flag	id_12_missing_flag	id_13_missing_flag	id_14_missing_flag	id_15_missing_flag	id_16_missing_flag	id_17_missing_flag	id_18_missing_flag	id_19_missing_flag	id_20_missing_flag	id_21_missing_flag	id_22_missing_flag	id_23_missing_flag	id_24_missing_flag	id_25_missing_flag	id_26_missing_flag	id_27_missing_flag	id_28_missing_flag	id_29_missing_flag	id_30_missing_flag	id_31_missing_flag	id_32_missing_flag	id_33_missing_flag	id_34_missing_flag	id_35_missing_flag	id_36_missing_flag	id_37_missing_flag	id_38_missing_flag	DeviceType_missing_flag	DeviceInfo_missing_flag	Date	_Weekdays	_Days	Trans_min_mean	Trans_min_std	TransactionAmt_to_mean_card1	TransactionAmt_to_mean_card4	TransactionAmt_to_std_card1	TransactionAmt_to_std_card4
0	2987000	86400	4.226834	W	13926	NaN	150.0	discover	142.0	credit	315.0	87.0	19.0	NoInf	NoInf	1.0	1.0	1.0	0.0	1.0	0.0	2.0	1.0	1.0	14.0	NaN	13.0	NaN	NaN	NaN	NaN	NaN	13.0	13.0	NaN	NaN	NaN	0.0	T	T	T	M2	F	T	NaN	NaN	NaN	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	1.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	117.0	0.0	0.0	0.0	0.0	117.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	...	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	True	True	True	True	True	True	True	True	False	False	True	False	False	False	False	True	False	False	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	2017-12-02 00:00:00	5	2	-66.527347	-0.278174	0.194638	0.257761	0.184560	0.170241
1	2987001	86401	3.367296	W	2755	404.0	150.0	mastercard	102.0	credit	325.0	87.0	NaN	Google	NoInf	1.0	1.0	1.0	0.0	0.0	0.0	1.0	1.0	1.0	0.0	NaN	NaN	0.0	NaN	NaN	NaN	NaN	0.0	NaN	NaN	NaN	NaN	0.0	NaN	NaN	NaN	M0	T	T	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0.0	0.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	...	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	True	True	False	False	True	True	True	True	False	False	False	False	False	False	False	True	False	False	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	2017-12-02 00:00:01	5	2	-106.027347	-0.443337	0.123780	0.219053	0.063004	0.114214
2	2987002	86469	4.077537	W	4663	490.0	150.0	visa	166.0	debit	330.0	87.0	287.0	Microsoft	NoInf	1.0	1.0	1.0	0.0	1.0	0.0	1.0	1.0	1.0	0.0	NaN	NaN	0.0	NaN	NaN	NaN	NaN	0.0	315.0	NaN	NaN	NaN	315.0	T	T	T	M0	F	F	F	F	F	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	...	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	True	True	False	False	False	False	False	True	False	False	False	True	False	False	True	True	True	True	True	True	True	False	False	True	False	True	True	True	False	False	False	False	False	False	2017-12-02 00:01:09	5	2	-76.027347	-0.317897	0.608151	0.443070	0.589226	0.258550
3	2987003	86499	3.912023	W	18132	567.0	150.0	mastercard	117.0	debit	476.0	87.0	NaN	Yahoo Mail	NoInf	2.0	5.0	4.0	0.0	1.0	0.0	1.0	25.0	1.0	112.0	112.0	0.0	94.0	0.0	NaN	NaN	NaN	84.0	NaN	NaN	NaN	NaN	111.0	NaN	NaN	NaN	M0	T	F	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	48.0	28.0	10.0	4.0	1.0	38.0	24.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	50.0	1758.0	925.0	354.0	135.0	50.0	1404.0	790.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	...	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	True	True	False	False	True	True	True	True	False	False	False	True	False	False	False	True	False	False	True	True	True	True	True	True	True	False	False	True	False	True	True	True	False	False	False	False	False	True	2017-12-02 00:01:39	5	2	-85.027347	-0.355529	0.405136	0.377678	0.259460	0.196921
4	2987004	86506	3.912023	H	4497	514.0	150.0	mastercard	102.0	credit	420.0	87.0	NaN	Google	NoInf	1.0	1.0	1.0	1.0	0.0	1.0	1.0	1.0	1.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	6.0	18.0	140.0	0.0	0.0	0.0	0.0	1803.0	49.0	64.0	0.0	0.0	0.0	0.0	0.0	0.0	15560.0	169690.796875	0.0	0.0	0.0	515.0	5155.0	2840.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	...	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	True	True	False	False	False	False	True	False	False	False	False	False	False	False	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	2017-12-02 00:01:46	5	2	-85.027347	-0.355529	0.515616	0.377678	0.882898	0.196921

	TransactionID	TransactionDT	TransactionAmt	ProductCD	card1	card2	card3	card4	card5	card6	addr1	addr2	dist1	P_emaildomain	R_emaildomain	C1	C2	C6	C8	C9	C10	C11	C13	C14	D1	D2	D3	D4	D5	D6	D8	D9	D10	D11	D12	D13	D14	D15	M1	M2	M3	M4	M5	M6	M7	M8	M9	id_01	id_02	id_03	id_04	id_05	id_06	id_09	id_10	id_11	id_12	id_13	id_14	id_15	id_16	id_17	id_19	id_20	id_28	id_29	id_30	id_31	id_32	id_33	id_34	id_35	id_36	id_37	id_38	DeviceType	DeviceInfo	card2_missing_flag	card3_missing_flag	card4_missing_flag	card5_missing_flag	card6_missing_flag	addr1_missing_flag	addr2_missing_flag	dist1_missing_flag	dist2_missing_flag	P_emaildomain_missing_flag	R_emaildomain_missing_flag	D1_missing_flag	D2_missing_flag	D3_missing_flag	D4_missing_flag	D5_missing_flag	D6_missing_flag	D7_missing_flag	D8_missing_flag	D9_missing_flag	D10_missing_flag	D11_missing_flag	D12_missing_flag	D13_missing_flag	D14_missing_flag	D15_missing_flag	M1_missing_flag	M2_missing_flag	M3_missing_flag	M4_missing_flag	M5_missing_flag	M6_missing_flag	M7_missing_flag	M8_missing_flag	M9_missing_flag	V1_missing_flag	V2_missing_flag	V3_missing_flag	V4_missing_flag	V5_missing_flag	V6_missing_flag	V7_missing_flag	V8_missing_flag	V9_missing_flag	V10_missing_flag	V11_missing_flag	V12_missing_flag	V13_missing_flag	V14_missing_flag	V15_missing_flag	V16_missing_flag	V17_missing_flag	V18_missing_flag	V19_missing_flag	V20_missing_flag	V21_missing_flag	V22_missing_flag	V23_missing_flag	V24_missing_flag	V25_missing_flag	V26_missing_flag	V27_missing_flag	V28_missing_flag	V29_missing_flag	V30_missing_flag	V31_missing_flag	V32_missing_flag	V33_missing_flag	V34_missing_flag	V35_missing_flag	V36_missing_flag	V37_missing_flag	V38_missing_flag	V39_missing_flag	V40_missing_flag	V41_missing_flag	V42_missing_flag	V43_missing_flag	V44_missing_flag	V45_missing_flag	V46_missing_flag	V47_missing_flag	V48_missing_flag	V49_missing_flag	V50_missing_flag	V51_missing_flag	V52_missing_flag	V53_missing_flag	V54_missing_flag	V55_missing_flag	V56_missing_flag	V57_missing_flag	V58_missing_flag	V59_missing_flag	V60_missing_flag	V61_missing_flag	V62_missing_flag	V63_missing_flag	V64_missing_flag	V65_missing_flag	V66_missing_flag	V67_missing_flag	V68_missing_flag	V69_missing_flag	V70_missing_flag	V71_missing_flag	V72_missing_flag	V73_missing_flag	V74_missing_flag	V75_missing_flag	V76_missing_flag	V77_missing_flag	V78_missing_flag	V79_missing_flag	V80_missing_flag	V81_missing_flag	V82_missing_flag	V83_missing_flag	V84_missing_flag	V85_missing_flag	V86_missing_flag	V87_missing_flag	V88_missing_flag	V89_missing_flag	V90_missing_flag	V91_missing_flag	V92_missing_flag	V93_missing_flag	V94_missing_flag	V95_missing_flag	V96_missing_flag	V97_missing_flag	V98_missing_flag	V99_missing_flag	V100_missing_flag	V101_missing_flag	V102_missing_flag	V103_missing_flag	V104_missing_flag	V105_missing_flag	V106_missing_flag	V107_missing_flag	V108_missing_flag	V109_missing_flag	V110_missing_flag	V111_missing_flag	V112_missing_flag	V113_missing_flag	V114_missing_flag	V115_missing_flag	V116_missing_flag	V117_missing_flag	V118_missing_flag	V119_missing_flag	V120_missing_flag	V121_missing_flag	V122_missing_flag	V123_missing_flag	V124_missing_flag	V125_missing_flag	V126_missing_flag	V127_missing_flag	V128_missing_flag	V129_missing_flag	V130_missing_flag	V131_missing_flag	V132_missing_flag	...	V170_missing_flag	V171_missing_flag	V172_missing_flag	V173_missing_flag	V174_missing_flag	V175_missing_flag	V176_missing_flag	V177_missing_flag	V178_missing_flag	V179_missing_flag	V180_missing_flag	V181_missing_flag	V182_missing_flag	V183_missing_flag	V184_missing_flag	V185_missing_flag	V186_missing_flag	V187_missing_flag	V188_missing_flag	V189_missing_flag	V190_missing_flag	V191_missing_flag	V192_missing_flag	V193_missing_flag	V194_missing_flag	V195_missing_flag	V196_missing_flag	V197_missing_flag	V198_missing_flag	V199_missing_flag	V200_missing_flag	V201_missing_flag	V202_missing_flag	V203_missing_flag	V204_missing_flag	V205_missing_flag	V206_missing_flag	V207_missing_flag	V208_missing_flag	V209_missing_flag	V210_missing_flag	V211_missing_flag	V212_missing_flag	V213_missing_flag	V214_missing_flag	V215_missing_flag	V216_missing_flag	V217_missing_flag	V218_missing_flag	V219_missing_flag	V220_missing_flag	V221_missing_flag	V222_missing_flag	V223_missing_flag	V224_missing_flag	V225_missing_flag	V226_missing_flag	V227_missing_flag	V228_missing_flag	V229_missing_flag	V230_missing_flag	V231_missing_flag	V232_missing_flag	V233_missing_flag	V234_missing_flag	V235_missing_flag	V236_missing_flag	V237_missing_flag	V238_missing_flag	V239_missing_flag	V240_missing_flag	V241_missing_flag	V242_missing_flag	V243_missing_flag	V244_missing_flag	V245_missing_flag	V246_missing_flag	V247_missing_flag	V248_missing_flag	V249_missing_flag	V250_missing_flag	V251_missing_flag	V252_missing_flag	V253_missing_flag	V254_missing_flag	V255_missing_flag	V256_missing_flag	V257_missing_flag	V258_missing_flag	V259_missing_flag	V260_missing_flag	V261_missing_flag	V262_missing_flag	V263_missing_flag	V264_missing_flag	V265_missing_flag	V266_missing_flag	V267_missing_flag	V268_missing_flag	V269_missing_flag	V270_missing_flag	V271_missing_flag	V272_missing_flag	V273_missing_flag	V274_missing_flag	V275_missing_flag	V276_missing_flag	V277_missing_flag	V278_missing_flag	V279_missing_flag	V280_missing_flag	V281_missing_flag	V282_missing_flag	V283_missing_flag	V284_missing_flag	V285_missing_flag	V286_missing_flag	V287_missing_flag	V288_missing_flag	V289_missing_flag	V290_missing_flag	V291_missing_flag	V292_missing_flag	V293_missing_flag	V294_missing_flag	V295_missing_flag	V296_missing_flag	V297_missing_flag	V298_missing_flag	V299_missing_flag	V300_missing_flag	V301_missing_flag	V302_missing_flag	V303_missing_flag	V304_missing_flag	V305_missing_flag	V306_missing_flag	V307_missing_flag	V308_missing_flag	V309_missing_flag	V310_missing_flag	V311_missing_flag	V312_missing_flag	V313_missing_flag	V314_missing_flag	V315_missing_flag	V316_missing_flag	V317_missing_flag	V318_missing_flag	V319_missing_flag	V320_missing_flag	V321_missing_flag	V322_missing_flag	V323_missing_flag	V324_missing_flag	V325_missing_flag	V326_missing_flag	V327_missing_flag	V328_missing_flag	V329_missing_flag	V330_missing_flag	V331_missing_flag	V332_missing_flag	V333_missing_flag	V334_missing_flag	V335_missing_flag	V336_missing_flag	V337_missing_flag	V338_missing_flag	V339_missing_flag	id_01_missing_flag	id_02_missing_flag	id_03_missing_flag	id_04_missing_flag	id_05_missing_flag	id_06_missing_flag	id_07_missing_flag	id_08_missing_flag	id_09_missing_flag	id_10_missing_flag	id_11_missing_flag	id_12_missing_flag	id_13_missing_flag	id_14_missing_flag	id_15_missing_flag	id_16_missing_flag	id_17_missing_flag	id_18_missing_flag	id_19_missing_flag	id_20_missing_flag	id_21_missing_flag	id_22_missing_flag	id_23_missing_flag	id_24_missing_flag	id_25_missing_flag	id_26_missing_flag	id_27_missing_flag	id_28_missing_flag	id_29_missing_flag	id_30_missing_flag	id_31_missing_flag	id_32_missing_flag	id_33_missing_flag	id_34_missing_flag	id_35_missing_flag	id_36_missing_flag	id_37_missing_flag	id_38_missing_flag	DeviceType_missing_flag	DeviceInfo_missing_flag	Date	_Weekdays	_Days	Trans_min_mean	Trans_min_std	TransactionAmt_to_mean_card1	TransactionAmt_to_mean_card4	TransactionAmt_to_std_card1	TransactionAmt_to_std_card4	PCA_V_0	PCA_V_1	PCA_V_2	PCA_V_3	PCA_V_4	PCA_V_5	PCA_V_6	PCA_V_7	PCA_V_8	PCA_V_9	PCA_V_10	PCA_V_11	PCA_V_12	PCA_V_13	PCA_V_14	PCA_V_15	PCA_V_16	PCA_V_17	PCA_V_18	PCA_V_19	PCA_V_20	PCA_V_21	PCA_V_22	PCA_V_23	PCA_V_24	PCA_V_25	PCA_V_26	PCA_V_27	PCA_V_28	PCA_V_29
0	2987000	86400	4.226562	W	13926	NaN	150.0	discover	142.0	credit	315.0	87.0	19.0	NoInf	NoInf	1.0	1.0	1.0	0.0	1.0	0.0	2.0	1.0	1.0	14.0	NaN	13.0	NaN	NaN	NaN	NaN	NaN	13.0	13.0	NaN	NaN	NaN	0.0	T	T	T	M2	F	T	NaN	NaN	NaN	0.0	70787.0	NaN	NaN	NaN	NaN	NaN	NaN	100.0	NotFound	NaN	-480.0	New	NotFound	166.0	542.0	144.0	New	NotFound	Android	Samsung	32.0	2220x1080	match_status:2	T	F	T	T	mobile	SAMSUNG SM-G892A Build/NRD90M	True	False	False	False	False	False	False	False	True	True	True	False	True	False	True	True	True	True	True	True	False	False	True	True	True	False	False	False	False	False	False	False	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	...	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	True	True	True	True	True	True	True	True	False	False	True	False	False	False	False	True	False	False	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	2017-12-02 00:00:00	5	2	-66.5	-0.278076	0.194580	0.257812	0.184560	0.170288	-0.157349	0.919434	-0.843750	0.308105	-0.089417	0.003044	-0.020050	-0.187622	0.038208	0.002604	-0.010536	0.034058	-0.044434	-0.089722	0.044769	0.001550	-0.003441	0.018616	-0.018387	0.010078	-0.026947	-0.021362	-0.054626	0.025375	0.018814	-0.006039	0.004055	-0.043335	0.008026	-0.007957
1	2987001	86401	3.367188	W	2755	404.0	150.0	mastercard	102.0	credit	325.0	87.0	NaN	Google	NoInf	1.0	1.0	1.0	0.0	0.0	0.0	1.0	1.0	1.0	0.0	NaN	NaN	0.0	NaN	NaN	NaN	NaN	0.0	NaN	NaN	NaN	NaN	0.0	NaN	NaN	NaN	M0	T	T	NaN	NaN	NaN	-5.0	98945.0	NaN	NaN	0.0	-5.0	NaN	NaN	100.0	NotFound	49.0	-300.0	New	NotFound	166.0	621.0	500.0	New	NotFound	iOS	Safari	32.0	1334x750	match_status:1	T	F	F	T	mobile	iOS Device	False	False	False	False	False	False	False	True	True	False	True	False	True	True	False	True	True	True	True	True	False	True	True	True	True	False	True	True	True	False	False	False	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	...	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	True	True	False	False	True	True	True	True	False	False	False	False	False	False	False	True	False	False	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	2017-12-02 00:00:01	5	2	-106.0	-0.443359	0.123779	0.218994	0.063004	0.114197	-0.086365	-0.800293	-0.152344	-0.363525	-0.101868	-0.002291	0.032318	-0.068848	0.040222	-0.180176	-0.059387	0.002302	0.018982	-0.029556	0.016647	-0.006241	-0.004208	0.010170	-0.001647	-0.022919	0.006298	-0.021164	0.054626	-0.042542	-0.026794	0.003531	0.001647	0.001576	-0.003611	-0.003761
2	2987002	86469	4.078125	W	4663	490.0	150.0	visa	166.0	debit	330.0	87.0	287.0	Microsoft	NoInf	1.0	1.0	1.0	0.0	1.0	0.0	1.0	1.0	1.0	0.0	NaN	NaN	0.0	NaN	NaN	NaN	NaN	0.0	315.0	NaN	NaN	NaN	315.0	T	T	T	M0	F	F	F	F	F	-5.0	191631.0	0.0	0.0	0.0	0.0	0.0	0.0	100.0	NotFound	52.0	NaN	Found	Found	121.0	410.0	142.0	Found	Found	NAN	Chrome	NaN	NaN	NaN	F	F	T	T	desktop	Windows	False	False	False	False	False	False	False	False	True	False	True	False	True	True	False	True	True	True	True	True	False	False	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	...	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	True	True	False	False	False	False	False	True	False	False	False	True	False	False	True	True	True	True	True	True	True	False	False	True	False	True	True	True	False	False	False	False	False	False	2017-12-02 00:01:09	5	2	-76.0	-0.317871	0.607910	0.443115	0.589226	0.258545	-0.800781	0.316895	0.273193	-0.026352	0.043182	-0.008064	-0.039276	-0.217041	0.017715	0.033508	-0.000322	-0.015343	0.020676	-0.046051	-0.006725	0.004875	0.001104	0.007484	-0.007793	-0.006611	-0.008270	0.009857	-0.007710	0.003191	0.002834	0.001886	0.003839	0.002903	-0.019592	-0.003424
3	2987003	86499	3.912109	W	18132	567.0	150.0	mastercard	117.0	debit	476.0	87.0	NaN	Yahoo Mail	NoInf	2.0	5.0	4.0	0.0	1.0	0.0	1.0	25.0	1.0	112.0	112.0	0.0	94.0	0.0	NaN	NaN	NaN	84.0	NaN	NaN	NaN	NaN	111.0	NaN	NaN	NaN	M0	T	F	NaN	NaN	NaN	-5.0	221832.0	NaN	NaN	0.0	-6.0	NaN	NaN	100.0	NotFound	52.0	NaN	New	NotFound	225.0	176.0	507.0	New	NotFound	NAN	Chrome	NaN	NaN	NaN	F	F	T	T	desktop	NaN	False	False	False	False	False	False	False	True	True	False	True	False	False	False	False	False	True	True	True	True	False	True	True	True	True	False	True	True	True	False	False	False	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	...	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	True	True	False	False	True	True	True	True	False	False	False	True	False	False	False	True	False	False	True	True	True	True	True	True	True	False	False	True	False	True	True	True	False	False	False	False	False	True	2017-12-02 00:01:39	5	2	-85.0	-0.355469	0.405029	0.377686	0.259460	0.196899	-0.237427	-0.811523	-0.123657	-0.423828	-0.067261	0.025040	0.110413	-0.253906	0.004803	0.170410	-0.012550	-0.014488	0.005268	0.031891	-0.013489	-0.017319	-0.001888	-0.084717	0.050293	0.140747	0.058960	-0.020218	0.066589	-0.010910	-0.017868	0.025528	0.003674	0.003511	0.026047	-0.041962
4	2987004	86506	3.912109	H	4497	514.0	150.0	mastercard	102.0	credit	420.0	87.0	NaN	Google	NoInf	1.0	1.0	1.0	1.0	0.0	1.0	1.0	1.0	1.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0.0	7460.0	0.0	0.0	1.0	0.0	0.0	0.0	100.0	NotFound	NaN	-300.0	Found	Found	166.0	529.0	575.0	Found	Found	Mac	Chrome	24.0	1280x800	match_status:2	T	F	T	T	desktop	MacOS	False	False	False	False	False	False	False	True	True	False	True	False	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	...	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	False	True	True	False	False	False	False	True	False	False	False	False	False	False	False	True	True	True	True	True	True	True	False	False	False	False	False	False	False	False	False	False	False	False	False	2017-12-02 00:01:46	5	2	-85.0	-0.355469	0.515625	0.377686	0.882898	0.196899	2.904297	0.380127	0.480713	-0.009659	-0.171753	1.170898	-0.178223	0.004517	0.043793	-0.001614	0.015610	-0.017883	-0.020523	-0.005604	-0.010262	-0.003822	-0.011459	-0.008179	0.013390	0.017792	-0.013702	0.000093	-0.005245	-0.142334	0.202271	0.014458	0.012764	0.002150	0.014008	-0.001770

	TransactionAmt	ProductCD	card1	card2	card3	card4	card5	card6	addr1	addr2	dist1	P_emaildomain	R_emaildomain	C1	C2	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	D1	D2	D3	D4	D5	D6	D8	D9	D10	D11	D12	D13	D14	D15	M1	M2	M3	M4	M5	M6	M7	M8	M9	id_01	id_02	id_03	id_04	id_05	id_06	id_09	id_10	id_11	id_12	id_13	id_14	id_15	id_16	id_17	id_19	id_20	id_28	id_29	id_30	id_31	id_32	id_33	id_34	id_35	id_36	id_37	id_38	DeviceType	DeviceInfo	addr1_missing_flag	addr2_missing_flag	dist1_missing_flag	dist2_missing_flag	R_emaildomain_missing_flag	D2_missing_flag	D3_missing_flag	D4_missing_flag	D5_missing_flag	D6_missing_flag	D7_missing_flag	D8_missing_flag	D9_missing_flag	D10_missing_flag	D11_missing_flag	D12_missing_flag	D13_missing_flag	D14_missing_flag	D15_missing_flag	M1_missing_flag	M2_missing_flag	M3_missing_flag	M4_missing_flag	M5_missing_flag	M6_missing_flag	M7_missing_flag	M8_missing_flag	M9_missing_flag	V1_missing_flag	V2_missing_flag	V3_missing_flag	V4_missing_flag	V5_missing_flag	V6_missing_flag	V7_missing_flag	V8_missing_flag	V9_missing_flag	V10_missing_flag	V11_missing_flag	V12_missing_flag	V13_missing_flag	V14_missing_flag	V15_missing_flag	V16_missing_flag	V17_missing_flag	V18_missing_flag	V19_missing_flag	V20_missing_flag	V21_missing_flag	V22_missing_flag	V23_missing_flag	V24_missing_flag	V25_missing_flag	V26_missing_flag	V27_missing_flag	V28_missing_flag	V29_missing_flag	V30_missing_flag	V31_missing_flag	V32_missing_flag	V33_missing_flag	V34_missing_flag	V35_missing_flag	V36_missing_flag	V37_missing_flag	V38_missing_flag	V39_missing_flag	V40_missing_flag	V41_missing_flag	V42_missing_flag	V43_missing_flag	V44_missing_flag	V45_missing_flag	V46_missing_flag	V47_missing_flag	V48_missing_flag	V49_missing_flag	V50_missing_flag	V51_missing_flag	V52_missing_flag	V53_missing_flag	V54_missing_flag	V55_missing_flag	V56_missing_flag	V57_missing_flag	V58_missing_flag	V59_missing_flag	V60_missing_flag	V61_missing_flag	V62_missing_flag	V63_missing_flag	V64_missing_flag	V65_missing_flag	V66_missing_flag	V67_missing_flag	V68_missing_flag	V69_missing_flag	V70_missing_flag	V71_missing_flag	V72_missing_flag	V73_missing_flag	V74_missing_flag	V75_missing_flag	V76_missing_flag	V77_missing_flag	V78_missing_flag	V79_missing_flag	V80_missing_flag	V81_missing_flag	V82_missing_flag	V83_missing_flag	V84_missing_flag	V85_missing_flag	V86_missing_flag	V87_missing_flag	V88_missing_flag	V89_missing_flag	V90_missing_flag	V91_missing_flag	V92_missing_flag	V93_missing_flag	V94_missing_flag	...	V169_missing_flag	V170_missing_flag	V171_missing_flag	V172_missing_flag	V173_missing_flag	V174_missing_flag	V175_missing_flag	V176_missing_flag	V177_missing_flag	V178_missing_flag	V179_missing_flag	V180_missing_flag	V181_missing_flag	V182_missing_flag	V183_missing_flag	V184_missing_flag	V185_missing_flag	V186_missing_flag	V187_missing_flag	V188_missing_flag	V189_missing_flag	V190_missing_flag	V191_missing_flag	V192_missing_flag	V193_missing_flag	V194_missing_flag	V195_missing_flag	V196_missing_flag	V197_missing_flag	V198_missing_flag	V199_missing_flag	V200_missing_flag	V201_missing_flag	V202_missing_flag	V203_missing_flag	V204_missing_flag	V205_missing_flag	V206_missing_flag	V207_missing_flag	V208_missing_flag	V209_missing_flag	V210_missing_flag	V211_missing_flag	V212_missing_flag	V213_missing_flag	V214_missing_flag	V215_missing_flag	V216_missing_flag	V217_missing_flag	V218_missing_flag	V219_missing_flag	V220_missing_flag	V221_missing_flag	V222_missing_flag	V223_missing_flag	V224_missing_flag	V225_missing_flag	V226_missing_flag	V227_missing_flag	V228_missing_flag	V229_missing_flag	V230_missing_flag	V231_missing_flag	V232_missing_flag	V233_missing_flag	V234_missing_flag	V235_missing_flag	V236_missing_flag	V237_missing_flag	V238_missing_flag	V239_missing_flag	V240_missing_flag	V241_missing_flag	V242_missing_flag	V243_missing_flag	V244_missing_flag	V245_missing_flag	V246_missing_flag	V247_missing_flag	V248_missing_flag	V249_missing_flag	V250_missing_flag	V251_missing_flag	V252_missing_flag	V253_missing_flag	V254_missing_flag	V255_missing_flag	V256_missing_flag	V257_missing_flag	V258_missing_flag	V259_missing_flag	V260_missing_flag	V261_missing_flag	V262_missing_flag	V263_missing_flag	V264_missing_flag	V265_missing_flag	V266_missing_flag	V267_missing_flag	V268_missing_flag	V269_missing_flag	V270_missing_flag	V271_missing_flag	V272_missing_flag	V273_missing_flag	V274_missing_flag	V275_missing_flag	V276_missing_flag	V277_missing_flag	V278_missing_flag	V322_missing_flag	V323_missing_flag	V324_missing_flag	V325_missing_flag	V326_missing_flag	V327_missing_flag	V328_missing_flag	V329_missing_flag	V330_missing_flag	V331_missing_flag	V332_missing_flag	V333_missing_flag	V334_missing_flag	V335_missing_flag	V336_missing_flag	V337_missing_flag	V338_missing_flag	V339_missing_flag	id_01_missing_flag	id_02_missing_flag	id_03_missing_flag	id_04_missing_flag	id_05_missing_flag	id_06_missing_flag	id_07_missing_flag	id_08_missing_flag	id_09_missing_flag	id_10_missing_flag	id_11_missing_flag	id_12_missing_flag	id_13_missing_flag	id_14_missing_flag	id_15_missing_flag	id_16_missing_flag	id_17_missing_flag	id_18_missing_flag	id_19_missing_flag	id_20_missing_flag	id_21_missing_flag	id_22_missing_flag	id_23_missing_flag	id_24_missing_flag	id_25_missing_flag	id_26_missing_flag	id_27_missing_flag	id_28_missing_flag	id_29_missing_flag	id_30_missing_flag	id_31_missing_flag	id_32_missing_flag	id_33_missing_flag	id_34_missing_flag	id_35_missing_flag	id_36_missing_flag	id_37_missing_flag	id_38_missing_flag	DeviceType_missing_flag	DeviceInfo_missing_flag	_Weekdays	_Hours	_Days	Trans_min_mean	Trans_min_std	TransactionAmt_to_mean_card1	TransactionAmt_to_mean_card4	TransactionAmt_to_std_card1	TransactionAmt_to_std_card4	PCA_V_0	PCA_V_1	PCA_V_2	PCA_V_3	PCA_V_4	PCA_V_5	PCA_V_6	PCA_V_7	PCA_V_8	PCA_V_9	PCA_V_10	PCA_V_11	PCA_V_12	PCA_V_13	PCA_V_14	PCA_V_15	PCA_V_16	PCA_V_17	PCA_V_18	PCA_V_19	PCA_V_20	PCA_V_21	PCA_V_22	PCA_V_23	PCA_V_24	PCA_V_25	PCA_V_26	PCA_V_27	PCA_V_28	PCA_V_29
464452	3.432	4	15063	514.0	150.0	4	226.0	1	315.0	87.0	18.0	0	2	1.0	1.0	0.0	1.0	1.0	0.0	0.0	1.0	0.0	1.0	0.0	2.0	1.0	35.0	35.0	34.0	34.0	34.0	NaN	NaN	NaN	34.0	34.0	NaN	NaN	NaN	34.0	1	1	1	3	2	1	0	0	1	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2	NaN	NaN	3	2	NaN	NaN	NaN	2	2	3	4	NaN	NaN	4	2	2	2	2	2	NaN	0	0	0	1	1	0	0	0	0	1	1	1	1	0	0	1	1	1	0	0	0	0	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	2	4	18	-104.06	-0.4350	0.1501	0.2324	0.092611	0.13560	-0.9330	0.3564	0.2969	-0.05145	0.012596	0.01012	0.02788	0.1576	-0.053300	0.13940	0.033940	0.021680	-0.04034	0.02574	-0.022110	-0.003386	0.019290	0.013780	-0.007637	-0.031980	0.012856	-0.007412	0.000978	-0.016700	-0.007600	-0.027450	0.00541	-0.006160	0.005920	0.008170
36372	3.717	0	6019	583.0	150.0	4	226.0	1	NaN	NaN	NaN	0	0	928.0	1301.0	475.0	0.0	475.0	476.0	688.0	0.0	823.0	667.0	667.0	695.0	300.0	0.0	NaN	NaN	0.0	NaN	0.0	NaN	NaN	NaN	NaN	0.0	0.0	302.0	0.0	2	2	2	2	2	2	2	2	2	-5.0	49333.0	0.0	0.0	0.0	-1.0	0.0	0.0	100.0	1	52.0	-360.0	0	0	166.0	410.0	611.0	0	0	4	0	24.0	0.02858	3	1	0	1	0	0	0.0808	1	1	1	0	0	1	1	0	1	0	1	1	1	1	1	0	0	0	0	1	1	1	0	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	0	0	0	0	0	1	1	0	0	0	0	0	0	0	0	0	0	0	0	1	1	1	1	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	5	11	-93.94	-0.3928	0.1823	0.3086	0.103778	0.18000	0.6445	-0.9595	0.6094	0.50200	-0.377700	-0.15760	0.07880	0.1742	1.205000	0.12660	-0.007683	-0.130900	-0.04294	0.00345	0.057530	0.048770	-0.015470	-0.016300	0.073850	0.016130	-0.018260	0.036070	-0.019200	-0.050660	-0.045530	-0.032350	-0.02182	0.038240	-0.009430	0.004864
572387	5.527	4	17188	321.0	150.0	4	226.0	2	299.0	87.0	1.0	6	2	1.0	1.0	0.0	0.0	1.0	0.0	0.0	2.0	0.0	1.0	0.0	14.0	1.0	568.0	80.0	1.0	569.0	1.0	NaN	NaN	NaN	569.0	569.0	NaN	NaN	NaN	569.0	1	1	1	1	0	1	0	1	1	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2	NaN	NaN	3	2	NaN	NaN	NaN	2	2	3	4	NaN	NaN	4	2	2	2	2	2	NaN	0	0	0	1	1	0	0	0	0	1	1	1	1	0	0	1	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	4	21	25	116.00	0.4849	1.9370	1.8850	1.348902	1.10000	-0.8013	0.3174	0.2720	-0.02687	0.049070	-0.00634	-0.02972	-0.2195	0.019070	0.03270	-0.009254	-0.009834	0.02269	-0.02753	-0.009140	-0.002127	-0.001934	-0.025130	0.016200	0.040620	0.000128	0.001646	-0.004090	0.010020	0.002650	0.010440	0.00992	0.005665	-0.008720	-0.011665
497276	3.950	4	1214	174.0	150.0	4	226.0	1	204.0	87.0	NaN	4	2	1.0	1.0	0.0	1.0	1.0	0.0	0.0	1.0	0.0	1.0	0.0	1.0	1.0	0.0	NaN	NaN	0.0	NaN	NaN	NaN	NaN	0.0	0.0	NaN	NaN	NaN	0.0	1	1	1	3	2	0	0	0	1	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2	NaN	NaN	3	2	NaN	NaN	NaN	2	2	3	4	NaN	NaN	4	2	2	2	2	2	NaN	0	0	1	1	1	1	1	0	1	1	1	1	1	0	0	1	1	1	0	0	0	0	1	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	0	19	30	-83.06	-0.3474	0.2540	0.3901	0.147945	0.22770	-0.9326	0.3564	0.2974	-0.05154	0.010610	0.00979	0.02590	0.1611	-0.054500	0.13270	0.035740	-0.010574	-0.00642	-0.03049	-0.012240	0.017960	0.018100	0.002857	-0.016740	0.007100	0.018840	-0.006878	0.008800	-0.004406	0.000245	-0.014570	-0.00884	-0.011110	-0.000597	-0.007458
97470	3.219	1	12695	490.0	150.0	4	226.0	2	325.0	87.0	NaN	6	6	1.0	1.0	1.0	0.0	1.0	0.0	1.0	0.0	1.0	1.0	0.0	1.0	1.0	0.0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	2	2	2	3	2	2	2	2	2	-5.0	68281.0	NaN	NaN	0.0	0.0	NaN	NaN	100.0	1	52.0	NaN	1	1	225.0	266.0	507.0	1	1	3	0	NaN	NaN	4	0	0	1	0	0	NaN	0	0	1	1	0	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	1	0	0	1	1	1	1	0	0	0	1	0	0	0	1	0	0	1	1	1	1	1	1	1	0	0	1	0	1	1	1	0	0	0	0	0	1	5	21	23	-110.00	-0.4600	0.1771	0.1877	0.116133	0.10956	2.8070	0.3457	0.3772	-0.01767	-0.042420	-0.24430	0.02310	0.0118	0.000914	0.02304	-0.009550	-0.022300	-0.02167	-0.01366	-0.005627	-0.009690	0.001454	-0.008240	-0.008900	0.005936	-0.004475	-0.000716	-0.003060	0.008514	-0.017840	0.000079	0.01335	-0.005287	-0.000648	0.007446

	y_test	y_pred	y_pred_prob	bin_y_pred_prob
7681	False	False	0.000884	(-0.0009859, 0.00121]
570242	False	False	0.012820	(0.00914, 0.0132]
340470	False	False	0.003954	(0.00333, 0.00477]
131781	False	False	0.008104	(0.00659, 0.00914]
472772	False	False	0.002139	(0.00121, 0.00219]

	prob_bin	not_fraud	fraud	perc_fraud	perc_not_fraud	cum_perc_fraud	cum_perc_not_fraud
0	Bin1((0.0529, 0.996])	12911	4806	0.785166	0.075485	78.516582	7.548483
1	Bin2((0.0253, 0.0529])	17203	513	0.083810	0.100578	86.897566	17.606305
2	Bin3((0.0157, 0.0253])	17447	269	0.043947	0.102005	91.292273	27.806783
3	Bin4((0.0111, 0.0157])	17549	167	0.027283	0.102601	94.020585	38.066896
4	Bin5((0.00852, 0.0111])	17589	127	0.020748	0.102835	96.095409	48.350396
5	Bin6((0.00665, 0.00852])	17629	87	0.014213	0.103069	97.516746	58.657281
6	Bin7((0.00507, 0.00665])	17669	47	0.007678	0.103303	98.284594	68.987553
7	Bin8((0.00374, 0.00507])	17673	43	0.007025	0.103326	98.987094	79.320163
8	Bin9((0.00266, 0.00374])	17682	34	0.005555	0.103379	99.542558	89.658035
9	Bin10((-0.000701, 0.00266])	17689	28	0.004574	0.103420	100.000000	100.000000

	prob_bin	not_fraud	fraud	perc_fraud	perc_not_fraud	cum_perc_fraud	cum_perc_not_fraud
0	Bin1((0.0532, 0.999])	13669	4048	0.661330	0.079917	66.132985	7.991651
1	Bin2((0.0311, 0.0532])	17001	715	0.116811	0.099397	77.814083	17.931373
2	Bin3((0.0213, 0.0311])	17318	398	0.065022	0.101251	84.316288	28.056431
3	Bin4((0.016, 0.0213])	17441	275	0.044927	0.101970	88.809018	38.253401
4	Bin5((0.013, 0.016])	17504	182	0.029734	0.102338	91.782388	48.487205
5	Bin6((0.0114, 0.013])	17574	172	0.028100	0.102747	94.592387	58.761934
6	Bin7((0.0097, 0.0114])	17552	109	0.017808	0.102619	96.373142	69.023801
7	Bin8((0.00801, 0.0097])	17682	88	0.014377	0.103379	97.810815	79.361674
8	Bin9((0.0068, 0.00801])	17636	81	0.013233	0.103110	99.134128	89.672652
9	Bin10((0.00179, 0.0068])	17664	53	0.008659	0.103273	100.000000	100.000000

	prob_bin	not_fraud	fraud	perc_fraud	perc_not_fraud	cum_perc_fraud	cum_perc_not_fraud
0	Bin1((0.587, 0.998])	13072	4645	0.758863	0.076426	75.886293	7.642612
1	Bin2((0.386, 0.587])	17080	636	0.103905	0.099859	86.276752	17.628522
2	Bin3((0.269, 0.386])	17398	318	0.051952	0.101718	91.471982	27.800352
3	Bin4((0.204, 0.269])	17520	196	0.032021	0.102432	94.674073	38.043510
4	Bin5((0.16, 0.204])	17593	123	0.020095	0.102858	96.683548	48.329348
5	Bin6((0.126, 0.16])	17650	66	0.010783	0.103192	97.761804	58.648511
6	Bin7((0.0985, 0.126])	17659	57	0.009312	0.103244	98.693024	68.972936
7	Bin8((0.0746, 0.0985])	17681	35	0.005718	0.103373	99.264826	79.310224
8	Bin9((0.0519, 0.0746])	17686	30	0.004901	0.103402	99.754942	89.650435
9	Bin10((0.0035399999999999997, 0.0519])	17702	15	0.002451	0.103496	100.000000	100.000000

	prob_bin	not_fraud	fraud	perc_fraud	perc_not_fraud	cum_perc_fraud	cum_perc_not_fraud
0	Bin1((0.71, 0.998])	13398	4319	0.705604	0.078332	70.560366	7.833210
1	Bin2((0.592, 0.71])	16984	732	0.119588	0.099298	82.519196	17.762992
2	Bin3((0.502, 0.592])	17337	379	0.061918	0.101362	88.710995	27.899159
3	Bin4((0.42, 0.502])	17452	263	0.042967	0.102034	93.007678	38.102560
4	Bin5((0.345, 0.42])	17566	151	0.024669	0.102701	95.474596	48.372612
5	Bin6((0.276, 0.345])	17603	113	0.018461	0.102917	97.320699	58.664297
6	Bin7((0.212, 0.276])	17643	73	0.011926	0.103151	98.513315	68.979368
7	Bin8((0.155, 0.212])	17674	42	0.006862	0.103332	99.199477	79.312562
8	Bin9((0.102, 0.155])	17680	36	0.005881	0.103367	99.787616	89.649265
9	Bin10((0.00214, 0.102])	17704	13	0.002124	0.103507	100.000000	100.000000

	prob_bin	not_fraud	fraud	perc_fraud	perc_not_fraud	cum_perc_fraud	cum_perc_not_fraud
0	Bin1((0.0513, 0.998])	13138	4579	0.748080	0.076812	74.808038	7.681199
1	Bin2((0.0278, 0.0513])	17093	623	0.101781	0.099935	84.986113	17.674710
2	Bin3((0.0217, 0.0278])	17408	308	0.050319	0.101777	90.017971	27.852386
3	Bin4((0.0189, 0.0217])	17490	226	0.036922	0.102256	93.710178	38.078005
4	Bin5((0.0172, 0.0189])	17598	118	0.019278	0.102888	95.637968	48.366766
5	Bin6((0.016, 0.0172])	17629	87	0.014213	0.103069	97.059304	58.673651
6	Bin7((0.0152, 0.016])	17649	67	0.010946	0.103186	98.153896	68.992230
7	Bin8((0.0145, 0.0152])	17664	52	0.008495	0.103273	99.003431	79.319578
8	Bin9((0.0138, 0.0145])	17678	38	0.006208	0.103355	99.624244	89.655112
9	Bin10((0.011800000000000001, 0.0138])	17694	23	0.003758	0.103449	100.000000	100.000000

	TransactionAmt	ProductCD	card1	card2	card3	card4	card5	card6	addr1	addr2	dist1	P_emaildomain	R_emaildomain	C1	C2	C4	C5	C6	C7	C8	C9	C10	C11	C12	C13	C14	D1	D2	D3	D4	D5	D6	D8	D9	D10	D11	D12	D13	D14	D15	M1	M2	M3	M4	M5	M6	M7	M8	M9	id_01	id_02	id_03	id_04	id_05	id_06	id_09	id_10	id_11	id_12	id_13	id_14	id_15	id_16	id_17	id_19	id_20	id_28	id_29	id_30	id_31	id_32	id_33	id_34	id_35	id_36	id_37	id_38	DeviceType	DeviceInfo	addr1_missing_flag	addr2_missing_flag	dist1_missing_flag	dist2_missing_flag	R_emaildomain_missing_flag	D2_missing_flag	D3_missing_flag	D4_missing_flag	D5_missing_flag	D6_missing_flag	D7_missing_flag	D8_missing_flag	D9_missing_flag	D10_missing_flag	D11_missing_flag	D12_missing_flag	D13_missing_flag	D14_missing_flag	D15_missing_flag	M1_missing_flag	M2_missing_flag	M3_missing_flag	M4_missing_flag	M5_missing_flag	M6_missing_flag	M7_missing_flag	M8_missing_flag	M9_missing_flag	V1_missing_flag	V2_missing_flag	V3_missing_flag	V4_missing_flag	V5_missing_flag	V6_missing_flag	V7_missing_flag	V8_missing_flag	V9_missing_flag	V10_missing_flag	V11_missing_flag	V12_missing_flag	V13_missing_flag	V14_missing_flag	V15_missing_flag	V16_missing_flag	V17_missing_flag	V18_missing_flag	V19_missing_flag	V20_missing_flag	V21_missing_flag	V22_missing_flag	V23_missing_flag	V24_missing_flag	V25_missing_flag	V26_missing_flag	V27_missing_flag	V28_missing_flag	V29_missing_flag	V30_missing_flag	V31_missing_flag	V32_missing_flag	V33_missing_flag	V34_missing_flag	V35_missing_flag	V36_missing_flag	V37_missing_flag	V38_missing_flag	V39_missing_flag	V40_missing_flag	V41_missing_flag	V42_missing_flag	V43_missing_flag	V44_missing_flag	V45_missing_flag	V46_missing_flag	V47_missing_flag	V48_missing_flag	V49_missing_flag	V50_missing_flag	V51_missing_flag	V52_missing_flag	V53_missing_flag	V54_missing_flag	V55_missing_flag	V56_missing_flag	V57_missing_flag	V58_missing_flag	V59_missing_flag	V60_missing_flag	V61_missing_flag	V62_missing_flag	V63_missing_flag	V64_missing_flag	V65_missing_flag	V66_missing_flag	V67_missing_flag	V68_missing_flag	V69_missing_flag	V70_missing_flag	V71_missing_flag	V72_missing_flag	V73_missing_flag	V74_missing_flag	V75_missing_flag	V76_missing_flag	V77_missing_flag	V78_missing_flag	V79_missing_flag	V80_missing_flag	V81_missing_flag	V82_missing_flag	V83_missing_flag	V84_missing_flag	V85_missing_flag	V86_missing_flag	V87_missing_flag	V88_missing_flag	V89_missing_flag	V90_missing_flag	V91_missing_flag	V92_missing_flag	V93_missing_flag	V94_missing_flag	...	V169_missing_flag	V170_missing_flag	V171_missing_flag	V172_missing_flag	V173_missing_flag	V174_missing_flag	V175_missing_flag	V176_missing_flag	V177_missing_flag	V178_missing_flag	V179_missing_flag	V180_missing_flag	V181_missing_flag	V182_missing_flag	V183_missing_flag	V184_missing_flag	V185_missing_flag	V186_missing_flag	V187_missing_flag	V188_missing_flag	V189_missing_flag	V190_missing_flag	V191_missing_flag	V192_missing_flag	V193_missing_flag	V194_missing_flag	V195_missing_flag	V196_missing_flag	V197_missing_flag	V198_missing_flag	V199_missing_flag	V200_missing_flag	V201_missing_flag	V202_missing_flag	V203_missing_flag	V204_missing_flag	V205_missing_flag	V206_missing_flag	V207_missing_flag	V208_missing_flag	V209_missing_flag	V210_missing_flag	V211_missing_flag	V212_missing_flag	V213_missing_flag	V214_missing_flag	V215_missing_flag	V216_missing_flag	V217_missing_flag	V218_missing_flag	V219_missing_flag	V220_missing_flag	V221_missing_flag	V222_missing_flag	V223_missing_flag	V224_missing_flag	V225_missing_flag	V226_missing_flag	V227_missing_flag	V228_missing_flag	V229_missing_flag	V230_missing_flag	V231_missing_flag	V232_missing_flag	V233_missing_flag	V234_missing_flag	V235_missing_flag	V236_missing_flag	V237_missing_flag	V238_missing_flag	V239_missing_flag	V240_missing_flag	V241_missing_flag	V242_missing_flag	V243_missing_flag	V244_missing_flag	V245_missing_flag	V246_missing_flag	V247_missing_flag	V248_missing_flag	V249_missing_flag	V250_missing_flag	V251_missing_flag	V252_missing_flag	V253_missing_flag	V254_missing_flag	V255_missing_flag	V256_missing_flag	V257_missing_flag	V258_missing_flag	V259_missing_flag	V260_missing_flag	V261_missing_flag	V262_missing_flag	V263_missing_flag	V264_missing_flag	V265_missing_flag	V266_missing_flag	V267_missing_flag	V268_missing_flag	V269_missing_flag	V270_missing_flag	V271_missing_flag	V272_missing_flag	V273_missing_flag	V274_missing_flag	V275_missing_flag	V276_missing_flag	V277_missing_flag	V278_missing_flag	V322_missing_flag	V323_missing_flag	V324_missing_flag	V325_missing_flag	V326_missing_flag	V327_missing_flag	V328_missing_flag	V329_missing_flag	V330_missing_flag	V331_missing_flag	V332_missing_flag	V333_missing_flag	V334_missing_flag	V335_missing_flag	V336_missing_flag	V337_missing_flag	V338_missing_flag	V339_missing_flag	id_01_missing_flag	id_02_missing_flag	id_03_missing_flag	id_04_missing_flag	id_05_missing_flag	id_06_missing_flag	id_07_missing_flag	id_08_missing_flag	id_09_missing_flag	id_10_missing_flag	id_11_missing_flag	id_12_missing_flag	id_13_missing_flag	id_14_missing_flag	id_15_missing_flag	id_16_missing_flag	id_17_missing_flag	id_18_missing_flag	id_19_missing_flag	id_20_missing_flag	id_21_missing_flag	id_22_missing_flag	id_23_missing_flag	id_24_missing_flag	id_25_missing_flag	id_26_missing_flag	id_27_missing_flag	id_28_missing_flag	id_29_missing_flag	id_30_missing_flag	id_31_missing_flag	id_32_missing_flag	id_33_missing_flag	id_34_missing_flag	id_35_missing_flag	id_36_missing_flag	id_37_missing_flag	id_38_missing_flag	DeviceType_missing_flag	DeviceInfo_missing_flag	_Weekdays	_Hours	_Days	Trans_min_mean	Trans_min_std	TransactionAmt_to_mean_card1	TransactionAmt_to_mean_card4	TransactionAmt_to_std_card1	TransactionAmt_to_std_card4	PCA_V_0	PCA_V_1	PCA_V_2	PCA_V_3	PCA_V_4	PCA_V_5	PCA_V_6	PCA_V_7	PCA_V_8	PCA_V_9	PCA_V_10	PCA_V_11	PCA_V_12	PCA_V_13	PCA_V_14	PCA_V_15	PCA_V_16	PCA_V_17	PCA_V_18	PCA_V_19	PCA_V_20	PCA_V_21	PCA_V_22	PCA_V_23	PCA_V_24	PCA_V_25	PCA_V_26	PCA_V_27	PCA_V_28	PCA_V_29
0	3.432	4.0	15063.0	514.0	150.0	4.0	226.0	1.0	315.00000	87.000000	18.00000	0.0	2.0	1.0	1.0	0.0	1.0	1.0	0.0	0.0	1.0	0.0	1.0	0.0	2.0	1.0	35.0	35.000000	34.000000	34.000000	34.000000	69.466886	145.896923	0.562258	34.000000	34.00000	53.870766	17.999492	57.910973	34.000000	1.0	1.0	1.0	3.0	2.0	1.0	0.0	0.0	1.0	-10.167535	175100.840652	0.063532	-0.055814	1.628387	-6.764911	0.095223	-0.30179	99.745253	2.0	48.029567	-344.788868	3.0	2.0	189.486242	352.986677	403.654199	2.0	2.0	3.0	4.0	26.519707	0.011556	4.0	2.0	2.0	2.0	2.0	2.0	0.041219	0.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	2.0	4.0	18.0	-104.06	-0.4350	0.1501	0.2324	0.092611	0.13560	-0.9330	0.3564	0.2969	-0.05145	0.012596	0.01012	0.02788	0.1576	-0.053300	0.13940	0.033940	0.021680	-0.04034	0.02574	-0.022110	-0.003386	0.019290	0.013780	-0.007637	-0.031980	0.012856	-0.007412	0.000978	-0.016700	-0.007600	-0.027450	0.00541	-0.006160	0.005920	0.008170
1	3.717	0.0	6019.0	583.0	150.0	4.0	226.0	1.0	290.73683	86.796065	119.26547	0.0	0.0	928.0	1301.0	475.0	0.0	475.0	476.0	688.0	0.0	823.0	667.0	667.0	695.0	300.0	0.0	169.661838	28.353969	0.000000	42.323807	0.000000	145.896923	0.562258	123.914942	146.43237	0.000000	0.000000	302.000000	0.000000	2.0	2.0	2.0	2.0	2.0	2.0	2.0	2.0	2.0	-5.000000	49333.000000	0.000000	0.000000	0.000000	-1.000000	0.000000	0.00000	100.000000	1.0	52.000000	-360.000000	0.0	0.0	166.000000	410.000000	611.000000	0.0	0.0	4.0	0.0	24.000000	0.028580	3.0	1.0	0.0	1.0	0.0	0.0	0.080800	1.0	1.0	1.0	0.0	0.0	1.0	1.0	0.0	1.0	0.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	5.0	11.0	-93.94	-0.3928	0.1823	0.3086	0.103778	0.18000	0.6445	-0.9595	0.6094	0.50200	-0.377700	-0.15760	0.07880	0.1742	1.205000	0.12660	-0.007683	-0.130900	-0.04294	0.00345	0.057530	0.048770	-0.015470	-0.016300	0.073850	0.016130	-0.018260	0.036070	-0.019200	-0.050660	-0.045530	-0.032350	-0.02182	0.038240	-0.009430	0.004864
2	5.527	4.0	17188.0	321.0	150.0	4.0	226.0	2.0	299.00000	87.000000	1.00000	6.0	2.0	1.0	1.0	0.0	0.0	1.0	0.0	0.0	2.0	0.0	1.0	0.0	14.0	1.0	568.0	80.000000	1.000000	569.000000	1.000000	69.466886	145.896923	0.562258	569.000000	569.00000	53.870766	17.999492	57.910973	569.000000	1.0	1.0	1.0	1.0	0.0	1.0	0.0	1.0	1.0	-10.167535	175100.840652	0.063532	-0.055814	1.628387	-6.764911	0.095223	-0.30179	99.745253	2.0	48.029567	-344.788868	3.0	2.0	189.486242	352.986677	403.654199	2.0	2.0	3.0	4.0	26.519707	0.011556	4.0	2.0	2.0	2.0	2.0	2.0	0.041219	0.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	4.0	21.0	25.0	116.00	0.4849	1.9370	1.8850	1.348902	1.10000	-0.8013	0.3174	0.2720	-0.02687	0.049070	-0.00634	-0.02972	-0.2195	0.019070	0.03270	-0.009254	-0.009834	0.02269	-0.02753	-0.009140	-0.002127	-0.001934	-0.025130	0.016200	0.040620	0.000128	0.001646	-0.004090	0.010020	0.002650	0.010440	0.00992	0.005665	-0.008720	-0.011665
3	3.950	4.0	1214.0	174.0	150.0	4.0	226.0	1.0	204.00000	87.000000	119.26547	4.0	2.0	1.0	1.0	0.0	1.0	1.0	0.0	0.0	1.0	0.0	1.0	0.0	1.0	1.0	0.0	169.661838	28.353969	0.000000	42.323807	69.466886	145.896923	0.562258	0.000000	0.00000	53.870766	17.999492	57.910973	0.000000	1.0	1.0	1.0	3.0	2.0	0.0	0.0	0.0	1.0	-10.167535	175100.840652	0.063532	-0.055814	1.628387	-6.764911	0.095223	-0.30179	99.745253	2.0	48.029567	-344.788868	3.0	2.0	189.486242	352.986677	403.654199	2.0	2.0	3.0	4.0	26.519707	0.011556	4.0	2.0	2.0	2.0	2.0	2.0	0.041219	0.0	0.0	1.0	1.0	1.0	1.0	1.0	0.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	19.0	30.0	-83.06	-0.3474	0.2540	0.3901	0.147945	0.22770	-0.9326	0.3564	0.2974	-0.05154	0.010610	0.00979	0.02590	0.1611	-0.054500	0.13270	0.035740	-0.010574	-0.00642	-0.03049	-0.012240	0.017960	0.018100	0.002857	-0.016740	0.007100	0.018840	-0.006878	0.008800	-0.004406	0.000245	-0.014570	-0.00884	-0.011110	-0.000597	-0.007458
4	3.219	1.0	12695.0	490.0	150.0	4.0	226.0	2.0	325.00000	87.000000	119.26547	6.0	6.0	1.0	1.0	1.0	0.0	1.0	0.0	1.0	0.0	1.0	1.0	0.0	1.0	1.0	0.0	169.661838	28.353969	139.771485	42.323807	69.466886	145.896923	0.562258	123.914942	146.43237	53.870766	17.999492	57.910973	163.670779	2.0	2.0	2.0	3.0	2.0	2.0	2.0	2.0	2.0	-5.000000	68281.000000	0.063532	-0.055814	0.000000	0.000000	0.095223	-0.30179	100.000000	1.0	52.000000	-344.788868	1.0	1.0	225.000000	266.000000	507.000000	1.0	1.0	3.0	0.0	26.519707	0.011556	4.0	0.0	0.0	1.0	0.0	0.0	0.041219	0.0	0.0	1.0	1.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	1.0	0.0	0.0	1.0	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	1.0	0.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	1.0	5.0	21.0	23.0	-110.00	-0.4600	0.1771	0.1877	0.116133	0.10956	2.8070	0.3457	0.3772	-0.01767	-0.042420	-0.24430	0.02310	0.0118	0.000914	0.02304	-0.009550	-0.022300	-0.02167	-0.01366	-0.005627	-0.009690	0.001454	-0.008240	-0.008900	0.005936	-0.004475	-0.000716	-0.003060	0.008514	-0.017840	0.000079	0.01335	-0.005287	-0.000648	0.007446

	prob_bin	not_fraud	fraud	perc_fraud	perc_not_fraud	cum_perc_fraud	cum_perc_not_fraud
0	Bin1((0.0194, 0.999])	12683	5034	0.822415	0.074152	82.241464	7.415181
1	Bin2((0.0146, 0.0194])	17276	440	0.071884	0.101005	89.429832	17.515683
2	Bin3((0.0134, 0.0146])	17484	232	0.037902	0.102221	93.220062	27.737794
3	Bin4((0.0128, 0.0134])	17576	140	0.022872	0.102759	95.507270	38.013693
4	Bin5((0.0125, 0.0128])	17620	96	0.015684	0.103016	97.075641	48.315316
5	Bin6((0.0122, 0.0125])	17654	62	0.010129	0.103215	98.088548	58.636818
6	Bin7((0.012, 0.0122])	17675	41	0.006698	0.103338	98.758373	68.970598
7	Bin8((0.0119, 0.012])	17681	35	0.005718	0.103373	99.330175	79.307885
8	Bin9((0.0117, 0.0119])	17686	30	0.004901	0.103402	99.820291	89.648096
9	Bin10((0.010499999999999999, 0.0117])	17706	11	0.001797	0.103519	100.000000	100.000000

MACHINE LEARNING WITH CALIBRATION TECHNIQUE TO DETECT FRAUDULENT CREDIT CARD TRANSACTIONS¶

Table Of Contents¶

1. Setup Environment ¶

2. Data Overview ¶

Identity Data Description¶

Transaction Data Description¶

3. Optimize Memory Used by Data ¶

Memory occupied by the dataframe (in mb)¶

4. Basic Data Stats ¶

Shape of dataframe¶

Summary of dataframe¶

Stats on Transaction Dataset¶

Inferences:

5. Data Preprocessing for EDA¶

Merge the datasets¶

Add missing flag¶

Clean Data¶

Create date features¶

6. Exploratory Data Analysis¶

Check distribution of target variable¶

Inferences:

Check distribution of Transaction Amount¶

Inferences:

Product Features¶

Inferences:

Card Features¶

Inferences:

Inferences:

P emaildomain¶

Inferences:

R-Email Domain¶

Inferences:

Days of the Month¶

Inferences:

Days of the week¶

Inferences:

Hour of the Day¶

Inferences:

Device Type¶

Inferences:

Columns from identity data¶

Get column names¶

7. Statistical Significance test ¶

Chi square test for categorical columns¶

Calculate odds¶

Odds¶

Odds Ratio¶

8. ANOVA Test¶

EDA Inferences:

Mini Challenge¶

7. Feature Engineering ¶

Domain Specific Features¶

8. Dimensionality Reduction - PCA ¶

9. Feature Encoding ¶

10. Data Preprocessing for Model Building ¶

11. Model Building ¶

12. XGBoost Classifier¶

13. Evaluation Metrics¶

14. Capture Rates and Calibration Curve ¶

Create validation set¶

Capture Rates Plot¶

Gains Table¶

Calibration Curve¶

Calibrate the model¶

XGBoost with booster = dart¶

Prediction¶

Evaluation Metrices¶

Inferences:

15. LightGBM ¶

Evaluation Metrices¶

Inferences:

16. Random Forest Classifier¶

17. Handling Class Imbalance ¶

Handle Class Imbalance with Random Oversampler¶

Evaluation Metrics¶

Inferences:

18. Cost Sensitive Learning with Class weights ¶

19. Model Calibration ¶

20. Model Tuning ¶

Evaluation Metrics¶

5. Data Preprocessing for EDA ¶

6. Exploratory Data Analysis ¶

13. Evaluation Metrics ¶

16. Random Forest Classifier ¶